2019 4th International Conference on Information Technology, Information Systems and Electrical Engineering (ICITISEE)

In this paper, we evaluate Optimum Golomb Ruler (OGR) for Next Generation Passive Optical Network Stage- 2 (NG-PON2) to overcome natural disaster namely non-linear effect. Non-Linear effect is the unavoidable tragedy from natural fiber effect, and biggest issue is Four Wave Mixing (FWM). To expand the capacity, NG-PON2 uses four wavelengths, where each wavelength transmit up to 10 Gbps for downstream. The impact of FWM decrease performance in NG-PON2, significantly, without exploit any mitigations. In addition, this paper focuses on the effect of the Optimum Unequal Channel Separation (OUCS) method with Golomb Ruler for the FWM effect on NG-PON2. Our simulation shows that after employee OGR, performance increase effectively at a distance of 25-31 km. The result shows that Q-Factor values, Signal to Noise Ratio (SNR), and received power improve around 14.27%, 40.26%, and 0.34%, respectively.

The phenomenon of mothers giving the wrong turn signal reminds us of the importance of giving a proper turn signal to other road users. Because the error gives the turn signal will have an impact on the occurrence of a massive accident risk. In the logistics business, it will never be separated from the mail carrier whose duty is to deliver packages. Duty as an introduction certainly has a high risk of accidents, besides having to be vigilant on the highway. The mail carrier must also think about the destination location of the package he will deliver so that the introduction is still looking for signs around to get to the location. This busyness, can result in incorrect giving of turn signal or not giving turn signal at all. So this study aims to analyze the possibility of making auto turn sign for post-delivery so that the delivery post can be safe and avoid the risk of accidents.

This paper presents the characterization of s-ring resonator as an electromagnetic wave absorber for C-band. The structure consists of copper s-ring patch in the top layer, FR-4 dielectric substrate and the back is fully copper laminated. The initial dimension has 7mm×7mm with the thickness of FR4 is 1mm. The characterization observes the changes of each variable from the initial design that affect the shift in resonant frequency and bandwidth. Maximum bandwidth can be achieved at S-Ring length 4.4mm for 0.0941 GHz.

This paper reports the performance of an indoor Visible Light Communication (VLC) system using a Direct Current-biased Optical Orthogonal Frequency Division Multiplexing (DCO-OFDM) scheme with a single Light Emitting Diode (LED). The DCO-OFDM is assigned to alter the bipolar signals produced by ordinary OFDM into unipolar signals. We simulate some M-ary modulations and employ Line-of-Sight (LOS) propagation model. The results show that the Quadrature Phase Shift Keying (QPSK) modulation is the most extensive one, with BER <= 10^-3 with lower SNR.

Today, many countries are increasing the energy mix of renewable energy, especially wind power, into the power system. This paper study the impact of high wind power penetration on the stability of the power system. The small signal stability (SSS) analysis of the doubly fed induction generator (DFIG) type wind power penetration was done by observing the power system's eigenvalue. Furthermore, the analysis of frequency oscillation, damping ratio, and participation factor was also done. In this research, a simulation to find out how the DFIG location and penetration level affect the power system stability was done in the IEEE 9-Bus system, using DIgSILENT/Power Factory. The simulation is done by replacing one synchronous generator with DFIG so that an investigation can be carried out to determine the effects of wind generator penetration. The result shows that wind power penetration has a negative impact on the power system's stability. It can be seen from the eigenvalue that goes to instability, the occurrence of local and inter-area oscillation, and the decrease in damping ratio of the system.

Machining of railway components needs to be monitored because this process requires precise accuracy and high surface finish of the final product. However, chatter vibration can be an obstacle which leads to negative effects. One of the ways for chatter monitoring in machining is by vibration signal analysis. In this paper, transient vibration signals are analyzed by spectral techniques; fast Fourier (FFT), short-time Fourier (STFT), and Hilbert-Huang (HHT) transforms. The results show the superiority of HHT for analysis transient signals in order to identify chatter vibration correctly in turning process

This study aimed to detect and count cacao pods on trees by taking outdoor images of cacao plants using a 4K resolution drone. The data acquisition was affected by the image acquisition distance and illumination in the plantation area that affected the objects'color surface. The data acquisition process is affected by different light illuminations that causing the color of the object's surface to be non-uniform and influencing the subsequent data processing. In this study, 85 images of cacao fruit on trees were taken with 3 variations of distances, which are 50 cm, 100 cm and 150 cm, and processed through several stages of data processing, which are image quality improvement using image enhancement at the preprocessing stage, image segmentation using the K-means method, and BLOB analysis method to detect and count the number of cocoa pods on the tree. The result shows an average accuracy of 93.3% at a distance of 50 cm.

In this paper we explore the recent phenomenon of influencer marketing in social media. Posts of the top-75 social media influencers on Instagram were analyzed over a 12-month period. The types of products in posts were identified for 226,801 images using an Inception V3 convolutional neural network (CNN). Products were then compared by level of engagement achieved to explore efficiency of engagement by product and influencer type. Results indicated that general influencers performed best, while travel influencers achieved greater overall engagement independent across product types, and specific types of influencers achieved better engagement for particular product types. The research points to the importance of fit between a product and influencer type in achieving impact via influencer marketing. Our findings offer help to brands in selecting influencers to endorse their products.

Instagram is one of the most popular social media platforms which is also commonly utilized to promote and market a product. One of the most efficient ways to market a product is by using endorsement scheme. However, nowadays a significant numbers of Instagram accounts have fake followers. Based on this problem, a system to support decision making on selecting which Instagram account is accountable for endorsement is essential. In this study, the authors used Simple Additive Weighting (SAW) method to analyze the parameters and Web Scraping to retrieve data from Instagram accounts automatically. Instagram account parameters observed in this study include the number of followers, number of average likes, number of comments, and average time interval between posts. The authors have successfully determined the best parameters to select accountable Instagram account for endorsement as shown by the results of the system's accuracy of 100%.

Boarding house are temporary homes that are most in demand by students or upper level students. However, to get the appropriate boarding house is not easy because of the difficulty of obtaining boarding house information and also many boarding house criteria that must be considered such as rental rates, room area and boarding house facilities etc. Multi criteria decision support system using the Topsis method can facilitate students in finding boarding house. Users can input the boarding house criteria to be searched such as rental rates, room area and boarding house facilities, then after that a mathematical calculation will be performed on each boarding house using the Topsis method. Resulting in board recommendations that match what the user is looking for. The system is made using Java for Android-based programming. The test is performed using the confusion matrix method to measure the accuracy of the costing recommendations produced by the system, resulting in 83% accuracy and 17% of recommendation errors.

The Fuzzy Multi Attribute Decision Making (FMADM) methods used in this study are FSAW and FTOPSIS. They are developed for the selection of Periodic Salary Increment (KGB) for the Civil Servants (PNS) of the East Kalimantan province because there is a big possibility of an error in entering data, calculating salary and calculating the time of submission in manually. The FMADM method selects and ranks employees according to qualifications for salary increases based on a number of criteria that refer to government regulations. Criteria that are used as references for salary increase selection of civil servants are includes: years of experience, assessment of Employee Performance Target (SKP) for the past two years, behavioral assessment, and disciplinary penalty. Based on the results of 40 employees data used in this study, the accuracy is as much as 90% compared with reality for FSAW, and the accuracy of FTOPSIS is as much as 85% from reality. The minimum preference threshold value is 0.70 to pass the Periodic Salary Increment selection

Public awareness of energy conservation can be realized through the process of electricity monitoring so that information can be found related to electricity usage and potential savings that can be made by the community. With the information on the results of this monitoring, the community can manage electricity usage more optimally so that it will have a positive impact on reducing electricity usage costs. The diversity of both linear and non-linear electrical equipment is a challenge in this monitoring process. In this study, the monitoring process is carried out in the aggregate on several electrical equipment indirectly on the electrical panel using the Non-Intrusive Load Monitoring (NILM) approach. The NILM approach is carried out through time and frequency domain analysis (TFA) to identify electrical equipment. Current signals (I) obtained in aggregate will be identified based on the characteristics of the frequency spectrum and its harmonics. The development of the TFA method was carried out to obtain feature vectors used for the identification process. Besides that, the noise reduction method was developed using the hybrid filter method (time domain and frequency domain) by combining Median Filter and Average FFT filter (MFAFFT). The use of hybrid filter is done to eliminate noise through the frequency domain that was previously still remaining when filtered using Median Filter in the time domain. Classification performance of the test shows better accuracy results on hybrid filtered data.

in a game, artificial intelligence plays a big role, a lot of logic and algorithms are applied therein, including in an RPG game. Fuzzy logic is a branch of artificial intelligence system (artificial intelligence) that emulates the ability of humans to think in the form of algorithms which are then run by machines. generally each opponent and player character has a health status as a life limit, the opponent character has a fuzzy input parameter is the length of the NPC life limit, distance to the player and health player. In this character also output parameters are blurred, dodged, attacked and calm. From the input parameters will then use fuzzy logic which will choose one output parameter. From the results of the implementation of producing activities when the conditions are weak and strong, the NPC is clever in acting in accordance with the logic of the conditions and as a final proof using MatLab simulation

This study is a follow-up study of the previous research researchers. In the previous research, a decision support system was built to select trainees in the BLK Bantul using the AHP-PROMETHEE method. But in this research there is no comparative analysis of the implementation of the AHP-PROMETHEE method with the AHP. Basically, the AHP method can be obtained alternative ranking without having to combine with methods outranking other. To research analysis was conducted implementation AHP comparison methods and AHP-PROMETHEE. The results of the comparison of the implementation of the AHP and AHP-PROMETHEE methods show the same results for the recommended participants accepted, only different sequences. The more appropriate method for resolving the problem of selecting trainees at the BLK Bantul is the AHP-PROMETHEE method.

Thesis is a paper that must be completed by students to meet the requirements of graduation. Students work on this paper in accordance with the topic and the concentration of their courses. In fact, many students who take thesis topics do not match the concentration of followed lectures. Therefore, it needs a decision support system to help students determine thesis topics. This system generates alternative decisions as an aid for deciding thesis topics. The process of making alternative decision uses a combination of SAW and TOPSIS methods. In addition, the determination of criteria weight is also influenced by K Means method. This clustering method serves to provide an alternative weighting value so that decision makers no longer need to give the initialization value. The combination of these methods generate a more absolute alternative value than only using SAW. This research is a development of previous research by adding TOPSIS method. The resulting alternative decision has a more absolute and significant alternative value. This is indicated by comparison of TOPSIS usage results and using only SAW.

Making learning method applying IT technology of choice for the learning process more interesting that combines augmented reality technology based on Android with educational book that discusses the process of metamorphosis. Aiming to create an atmosphere of teaching and learning to be more interactive, interesting, and the delivery of content clearer. Due to the 3D feature (three-dimensional) objects relating to the material, sound narration to clarify the material presented, and videos that explain more clearly of a process that is raised from the material and can be viewed live significantly through this technology. Learning media is running with a purpose, namely to Indonesia a better education. Maximize new learning method that is more effective than the application of learning methods previously applied. Creating a bond in the learning process between teachers and students and also between parent and child is more harmonious.

The use of Augmented Reality is an option in the delivery of information in the form of learning material because in Augmented Reality elements such as text, sound, images, and animation can be put together in its presentation, and are able to demonstrate a real object into a virtual form that can be seen from all sides. Based on the background above, the authors think that augmented reality technology can be used in conveying sun and earth material. this application will be used as a learning medium. the writer will test whether this application can be accepted as a learning medium for elementary school students

Noise cancellation is a common practice in audio signal processing that involves adaptive digital filter to achieve a clean audio signal. This paper presents the implementation of Least Means Squared (LMS) algorithm for real-time Active Noise Cancellation (ANC) by considering the surrounding sound signal that contains noise. This method was chosen to improve the output sound without damaging original sound using LMS adaptive digital filter by decreasing noise in decibel value. This work has been accomplished by the design and analyzes of LMS adaptive filter implemented on both MATLAB and Field Programmable Gate Array (FPGA). The design approach is done by developing the algorithm in MATLAB Simulink, then applied on FPGA hardware platform, Cyclone IV EP4CE115F29C7 board using Intel Quartus Prime development platform for simulation. Three major demonstrable hardware applications presented in this work are the real-time sound filter, ANC and sound measurement in decibel. Based on the standard sound level, 85dB (A) and not more than 115dB (A) was chosen for result analysis between filtered and unfiltered sound. In a comparison of three genre songs implementing the real-time ANC system containing LMS filter, the hardware implementation able to reduce the sound level at the range of 15.56% to 20.58%.

Content filtering application on internet is very important to protect children from various negative content such as pornography and violence. Most of the content is pornographic images, so it is necessary to find a detection tool for pornographic images that have high accuracy and relatively low computational time. The initial stage of detecting pornographic images is to be able to distinguish pornographic images from non-pornographic images by the classification method. To achieve these objectives, a comparative study of the existing classification methods such as K Nearest Neighbors (KNN), Logistic Regression (LR), Linear Discriminant Analysis (LDA), Decision Tree (DT), Random Forest (RF), is needed. Multi Layers Perceptron (MLP). To be able to do the classification needs to be carried out a feature extraction process that can distinguish pornographic images and non-pornographic images. In this study, the extraction method features Haralick texture (texture features), Color Histogram (color features), and Hu Moment (shape features). The test results show that the most suitable method for classification of pornographic images is the Random Forest (RF) with an accuracy of 91.04% and the computational time per image is 0.80 ms.

Lips are part of body parts in facial area which are unique in shapes and may have different patterns. Pattern or shape of lips can be used as an individual identity, either for a live or deceased person. Forensic science is a branch of medical science which utilizes lip shape to determine the identity of a person, including their sex. Determination of sex using image of lip shape using 2DPCA and SOM algorithms was successful in classifying the sex of a person, as male or female, using desktop-based Java programming language. This study employed 90 lip images, consisting of 60 training data (30 of male and 30 of female) and 30 testing data of male and female. .The testing result showed accuracy rate of 76,66% with a rate of 0,9 and maximum iteration of 10.000.

Copy move is one of the techniques used to fake images. Many studies have developed methods to detect displacement of counterfeit copies of images. Some researchers use methods for the process of extracting features in an image such as Scale Invariant Feature Transform and Speed Up Robust Feature. This study aims to compare the results of the detection of copy move image forgery using the Scale Invariant Feature Transform algorithm and Speed Up Robust Feature. Testing will be done is in terms of accuracy and execution time. At the pre-processing stage, the image decomposition process is done using the Discrete Wavelet Transform (DWT) method. The purpose of decomposition is to decompose the image into wavelet coefficients and scaling functions. In Discrete Wavelet Transform, signal energy concentrates to a certain wavelet coefficient. This characteristic is useful for compressing images. The result of the decommission that will be used for the next process is the LL sub-drawing. in this study there are 5 types of testing, namely without transformation, scale transformation, rotation transformation, lighting transformation, and blur transformation. Modeling the program code will use the MATLAB programming language. By using the transformation process, there are more key points, but fewer key points. Based on testing that has been done that the Scale Invariant Feature Transform algorithm has a higher accuracy than Speed Up Robust Feature. The difference in the 1st test was 16.32%, the difference in the 2nd test was 9.96%, the difference in the 3rd test was 13.32%, the difference in the 4th test was 13.44%, and the difference in the 2nd test 5 of 1.87 %. To test the execution time, the Speed Up Robust Feature algorithm has a faster time compared to the Scale Invariant Feature Transform. The difference in testing 1, 2, and 3 amounted to 1.97 seconds, the difference in the 4th test was 0.7 seconds and the difference in the fifth test was 0.62 seconds

Solar Lentigines is a skin disease caused by frequent exposure to direct sunlight. Appearance in solar lentigines can resemble Lentigo Malignant Melanoma cancer at an early stage. solar lentigines a disease that is not dangerous and does not require special treatment, but if there are significant changes such as asymmetrical wounds, obscure borders, non-homogeneous colors, diameters exceeding 6 millimeters, solar lentigines are suspected as lentigo malignant early stage melanoma. lentigo malignant melanoma is a rare but dangerous type of skin cancer if it is not treated immediately with asymmetrical, unclear boundaries, non-homogeneous colors, diameters exceeding 6 millimeters. this study aims to help detect the potential of lentigo malignant melanoma disease by using the image of solar lentigines. this application uses the feature extraction feature ABCD to scratch the input image. ABCD method is a medical method used to detect cancer in terms of asymmetry, obscure borders, color, diameter. The data of this study were obtained from one hospital in Bandung and the data was presented in table form and explained informally. The result of the application is a diagnosis of the potential for disease. The accuracy value of this application is 97.5% from 60 datasets.

Many evidence suggested that one's personality can be seen from his/her handwritten scratches, using Graphology analysis. Graphology consists of two techniques of structural and symbol analysis. Structural analysis is done by using variables such as margin, a space between lines, a space between words, slope, and dominant zone. While symbol analysis is based on how to write each letter. A computer-based graphology analysis examines the handwriting image and then gives out the result of one person's suggested personalities based on the pattern of each feature. This research conducted by using both techniques of six features. The multi-structure analysis was done to features of margin, space between lines, a space between words, slope, and dominant zone, while four specific letters ('a', 'g', 's', 't') were analyzed using the Convolutional Neural Networks (CNN) classification approach. The results showed that the accuracy of the structured approach was up to 82.5-100%, while the accuracy of the symbol approach using CNN had an accuracy of up to 98.03% of new data with 7-10 minutes in training process

Attending basic education is an obligation for all Indonesian citizens. The financial cost is one of input component to implement an education or even can be considered as the main requirement in achieving the goal of education. For a private education institution in Indonesia, financial cost is mainly covered from students' tuition payment. SMK Al-Islam Surakarta is a private school that manages all its students to pay school tuition fees monthly. According to its last year's administrative report, the number of students who are late in paying school tuition fee is around 60%. Since the school's operational costs are heavily depended on their income from tuition fees, this considered as an essential problem and has to be managed and predicted as well. This research will discuss about techniques in predicting the late payment of tuition fees. From many popular methods available in this area, we observed two of them namely Naïve Bayes and K-Nearest Neighbor (K-NN). This study will compare the accuracy between those two methods. The data used for the lab work is the official education basic data of Al-Islam Surakarta Vocational School in 2017/2018 totaling 236 data. In order to increase the accuracy, this study also combines the prediction methods with feature selection technique Information Gain which is commonly used to select optimal parameter for prediction process. In the end, the system is tested using the Confusion Matrix method. The results showed that the Naïve Bayes Method with Information Gain attribute selection produced the highest accuracy of 69%.

This paper evaluates the impact of device orientation in Visible Light Communication (VLC) with the Bit Error Rate (BER) as the main parameter of measurement. Most studies on optical wireless communications have neglected the effect of random orientation in their performance analysis. In this paper, we analyze the device orientation and assess its importance on system performance. The device orientation are we use are 0^o, 15^o, and 35^o. To support the simulation, we use the OOK-NRZ modulation techniques with the threshold Bit Error Rate (BER) around 10^-5. Each of the device orientation has the value of a wide range of communication coverage. The smaller of device orientation, the coverage will be wider. With using the biggest device orientation, the communication coverage is decreased of 11.04%, while the smallest device orientation is 98.08% of coverage area.

Optical fiber communication system is the most implemented technology for backbone and access network telecommunication. However, the mobility of users has forced technology wired to wireless technology and one of the newest research is Optical Wireless Communication (OWC). The OWC has no longer using fiber optic as propagation media and using air for its transmission. In this paper, we evaluate number of Light Emitting Diode (LED) with comparison for one and two transmitters towards light communication distribution in a 5 x 5 x 4 m on closed room with reflector. We also use reflector mirror in the wall and using On Off Keying-Non Return to Zero (OOK-NRZ) as a modulation system. In addition, we evaluate wall reflector for impact of communication coverage area. From computer simulation, the results show that two LEDs used with reflector has 29.9 % larger coverage area than one LED.

Smart home uses information technology to manage the devices inside the home to provide convenience for its owners. The most widely used technology for smart home systems is IoT. Nevertheless, IoT has challenges related to data security, moreover for the large scale and distributed IoT network properties. A solution that can be used to overcome this problem is by building a blockchain-based IoT system. The properties of blockchain which are immutable and non-repudiation can be used to store the smart home data. One of the data that needs to be secure is door lock access data. This data must not be vulnerable to counterfeiting and hacking, because a door is directly related to the safety of the homeowner. In this study, a blockchain-based data storage system is proposed for the door lock system. This study uses the Ethereum blockchain platform to store door lock access data and smart contracts to manage the policies. The proposed system is then tested using the avalanche effect to analyze the security level. The test results show the avalanche effect index for the proposed system is about 96% on average which means the proposed system can be categorized as a secure system.

Cigarette smoke can decrease air quality that affects human health. To create a smoke-free environment, it is important to have a system that can be used to detect the presence of hazardous gases in cigarette smoke. In this paper, we design and implement a wireless cigarette smoke detection system. The hardware utilises Arduino and four gas sensors, namely MQ-135, MQ-2, MQ-7 and MQ-9 to sense, collect, and send data wirelessly to Raspberry Pi using an ESP8266 ESP-01 Wi-Fi module. In the software side, we develop three programs, i.e. for Arduino, Raspberry Pi, and the web server. We show in real experiments that our system successfully gathers sensor readings from the gas sensors, sends data via wireless connection and stores it in the web server.

The MQTT protocol is an established bandwidth-efficient method of connection with remote locations using wireless sensor networks. It forms a backbone infrastructure for IoT collection and analytics. The collection and sharing of data by embedded devices and heterogeneous systems fuels an ever-increasing reliance on secure data. An important consideration is the resilience of MQTT brokers when subject to large subscriber loads. This research describes how the intelligence gathered by a simple physical reconnaissance attack can be used as part of a rapid kill-chain to facilitate resource starvation and water torture attacks in order to bring down three distinct MQTT broker implementations.

Internet of Things (IoT) model has limited interaction because devices must communicate within the same domain. In addition, if there is a new type of sensor relevant to a domain, the sensor cannot communicate because of domain differences. Because of these problems a system is needed to communicate between protocol domains. Communication between the multi-protocol domains of CoAP, MQTT and Websocket on the device can be achieved by creating a gateway for each protocol, then connecting with a broker. Each data from the sensor will be processed into ESP32 as a microcontroller. To overcome the problem of the differences in multi-protocol domains, namely by making a gateway on a one sensor that uses the MQTT Protocol and other sensors using the CoAP protocol. The sensor data is processed by the Raspberry Pi as a multi-protocol gateway. Then the data are multiplexed and be sent to the database server using the Web socket Protocol. From the testing result show that the CoAP protocol has a better performance than MQTT protocol. The delay of CoAP protocol achieved 28.80 ms and the delay of MQTT protocol is 39.59 ms.

Improved Technology until now started to help many aspects of human works, including medical facilities. One of them is Electroencephalography (EEG) which records electrical activities of the brain. This technology is used to diagnose any brain disease that makes an abnormality at EEG signal recording. EEG usually has non-invasive methods, where electrodes placed on the scalp that easy and reusable. This technology itself has been started in 1924 by Hans Berger. This paper discusses the basic construction of EEG, focused on simulation, device construction, and results from devices. It will further mention how Arduino Mega 2560 record analog signals, circuit diagrams that will be used, and final results that showed and processed on MatLab application. The final result would use Ten-Twenty Electrode System which can be compared with any result from any different recording and be displayed in MatLab with graph form that has been filtered with Fast Fourier Transform (FFT) results.

Issues of cogging torque that comes out on permanent magnet electric machines are crucial interests reduced, this is related to the performance of the machine itself. The purpose of this research is to diminish the cogging torque value so that the generator can perform optimally. The method employed to decline the cogging torque is an anti-notch method with stator tooth tip transformations in shape. There are 4 models, particularly no tip, tiered tip, rounded tip and parallel tip simulated using FEMM 4.2 software with two conditions, before and after using the anti-notch method. From the simulation result, it can be accomplished that by adding stator tooth tip the biggest cogging torque decline takes place in tiered tip model that is equal to 71.76%. Meanwhile, when added anti-notch to the surface of the stator, the biggest reduction in cogging torque takes place in the rounded tip model, which is diminished by 99.03% from the initial model.

In the actual environment, device-to-device (D2D) communication on an adhoc network experiences rapid signal quality changes. The variation in the amplitude and phase of the received signal is due to changes in the characteristics of the path over times which are affected by distance and multipath fading between the sender and receiver. This will affect the performance of routing connection in D2D communication. In this research, we investigated the impact of level crossing rate (LCR) on routing quality connection in adhoc network for D2D communication. LCR is quantifies how often the fading signal is below the threshold value which is usually to measure of the rapidity of the fading. The simulation result shows that higher LCR indicate low quality routing connection and reduce throughput performance. The multipath fading and the velocity lead higher level crossing rate in routing connection. Threshold signal level play important role in determining the level crossing rate and quality routing connection in path selection mechanism.

This paper presents an adaptive scaling factor for multiple watermarking scheme. Multiple watermarks are embedded in the red and green components with lowest variance of the cover image. DCT is applied to each selected block for transforming image pixels to DCT coefficients. This experiment examines DCT coefficients in the middle frequency to obtain adaptive scaling factor for embedding a watermark. The proposed algorithm used the impact of selected DCT coefficients to imperceptibility and robustness for generating t adaptive scaling factor. Arnold transform is used to improve the security and secrecy of the embedded watermark image. The experimental results reported that the proposed scheme gives better invisibility performance for multiple watermarks compared to existing schemes.

A modern city is characterized by many buildings for commercial and residential purposes. Disaster mitigation such as building fires, earthquakes, and so on should be taken into account in city management. Mitigation seeks to reduce the level of risks when a disaster occurs. A proper mitigation requires an early warning system to alert the city authority in order to handle a disaster effectively. This paper describes the prototype of a building monitoring system using an internet-of-things platform for supporting the disaster mitigation in the city. It employs a temperature sensor, a smoke sensor, and a vibration sensor to detect the condition of temperature, smoke concentration, and the presence or absence of vibrations as an indication of an earthquake respectively. All sensor data is pooled in XBee transmitter module to broadcast into XBee receiver module. The data is sent continuously through the ESP-8266 module such that an internet-of-things platform called Cayenne will proceed the data by displaying the condition of building and alerting a warning message. If there is an indication of a fire or earthquake, then such message will be displayed in the Cayenne dashboard and sent to the registered mobile phone number. Experimental results show the Cayenne dashboard can display data related to the condition of the building at any time and a warning message can be displayed whenever indications of a fire and earthquake.

MQTT is an open source message protocol ideally used in machine-to-machine or Internet of Things. MQTT designed for limited devices, low bandwidth, and high latency system. MQTT provided with transmission assurance by using the QoS level in publish and subscribe mechanism. The aim of this research is to check the MQTT performance as an implementation of monitoring and controlling system. The system used in this research designed to be able to automate a greenhouse prototyping system and also to monitor and control greenhouses remotely. The case of this system is to automatically control and monitor chili crops in the greenhouse prototyping. This system used Wemos D1 R2 board as a control center, SHT11 sensor, YL-69 sensor, 16x2 LCD, RTC, relay, and Level-0 QoS MQTT communication protocol. The parameter used to check the MQTT performance are delay, throughput, and packet loss. The result showed that MQTT protocol is suitable for transmitting data in Internet of Things system. The average QoS parameter result changed based on the transmission speed in each network and the use of level 0 QoS on MQTT protocol.

Automatic Waste Segregator and Monitoring System is an automated dustbin which can be used by all level of society. This paper proposes an Automatic Waste Segregator and Monitoring System which is designed to sort the wet waste, paper, plastic and aluminum material into each bin respectively. The main objective of this project is to automatically segregate the waste based on its material and monitor daily waste disposal rates and apply proper maintenance of the dustbin. Besides that, this project is to promote 3R (reuse, recycle, reduced) concept among citizens and prototype development that allows janitors to monitor the level of garbage in the dustbin detected by sensors inside the dustbin. When the bin reaches the threshold limit, the janitors is notified via SMS by using Icomsat 1.1 SIM900 GSM module. This project has been tested at the Faculty of Electrical Engineering UiTM Shah Alam. This dustbin will automatically open when someone is close to the dustbin and will not open if the garbage inside it has reached the threshold limit to prevent overflow and mess. Anyone who throws the rubbish in the dustbin will receive "thank you" on the LCD screen as a token of appreciation. In this paper, this project is built by using Microcontroller based platform Arduino Mega 2560 board which is interfaced with ESP8266 12E LoLin NodeMCU V3 module, Inductive proximity sensor, Laser module, Light Dependent Resistor (LDR), Liquid sensor module and Ultrasonic sensor.

The era of millimeter wave (mmWave) spectrum with operating frequency range 30-300 GHz is coming. These frequencies are potentially used by 5G cellular communication since it has some benefits such as larger bandwidth and capacity. Energy efficiency is a crucial issue in 5G mmWave where it employs a large number of antenna array with power consumed ADCs (Analog to Digital Converters). This paper studies the efficiency of energy for mmWave communication through the optimization of ADC bits quantization. We use Channel Structure-based Scheduling (CSS) algorithm for evaluation with the total rate used as a performance parameter. The selection of quantization bit must consider the complexity of implementation. The simulation uses 2 bits quantization to explore the total rate performance. Then increasing bit number into 3 bits and 5 bits in the following simulation. The total rate increases as transmit power increases for the entire implemented quantization bit number. The random scheduling algorithm is used as a reference for comparison. The simulation shows that Channel structure-based scheduling algorithm outperforms random scheduling.

Data exchange occurs with a variety of purposes, ranging from sharing information or just basic communication between people. Data exchanges that occur every day make data traffic busier. Data theft is made even easier by utilizing a large number of message senders and busy data traffic. One popular method used to secure data is the steganography method. In this research, the End of File steganography method is successfully implemented in various types of multimedia, namely image, audio, and video. The steganography process is done using a variety of secret message sizes, then some evaluations are carried out. In addition, the purpose of the evaluation is also to find out the comparative advantages between the types of media used. Evaluations that done for this research included changes in size, changes in physical cover files, histograms, and Mean of Score assessments. From all of the evaluations, it can be concluded that the most suitable multimedia files to be used as steganography media for the End of File method are the image and video media. However, this method provides the effect of increasing the size, so it is advisable to use video because generally, the video size is always larger than the image size because it contains images and audio in one file at a time.

Data security is an important aspect along huge amount of data is sent through the internet. Data needs to be secured and securing data can use cryptographic techniques. In symmetric cryptographic techniques, there are permutation and stream techniques, where each technique has its own weaknesses and strengths. This research combines the two techniques into hybrid techniques to get stronger encryption. The proposed permutation technique is cyclic bit shift, while the proposed stream technique is RC4. The results of the combination of the two methods are then performed performance tests using the avalanche effect (AE), bit error ratio (BER), the time required for decryption encryption and character error rate (CER). Based on the results of testing the proposed method is superior compared to the method that has been previously proposed.

Steganography is a method of embedding secret messages into a cover file in the form of text, audio, picture or video, so that the message is not suspected by those who are not authorized to open the message. The technique to find out whether the cover media is a stego file or not is steganalysis. In this study, detection of hidden messages focused on MP3 files inserted by the MP3Stego algorithm and Equal Length Entropy Codes Substitution to classify based on algorithms and the estimated length of the message and detect cover files. In conducting this research, it is necessary to know the audio features of MP3, build suitable deep learning methods and the performance of the models that have been produced. The proposed solution for these two problems is to use the QMDCT audio feature and deep learning architecture with Convolutional Neural Network. The results of this study are the best algorithm classification model with an accuracy performance of 91.78% and F1-Score 92.22% and the best classification model for message length estimation has an accuracy performance of 24.16% and F1-Score 21.40%. Thus, the proposal of deep learning architecture is good in classifying algorithms and covers, but still poor in classifying the estimated length of the message.

This paper presents an audio watermarking scheme that ensures the efficient of robustness-distortion ratio and a pleasing amount of capacity. Discrete Wavelet Transform (DWT synergism along with Discrete Cosine Transform (DCT), Singular Value Decomposition (SVD), and Cartesian to Polar Transformation (CPT) to emphasize copyright protection. The system establishes with DWT that divides the signal within a specific range of frequency, alter it to the frequency domain by DCT, partite with mathematical processing SVD, and transform the peculiar value of matrix-S to polar value using CPT, furthermore, QIM embed the watermark into the angular value from the CPT. The system performance is astonishing, tested using 4 different audio host signal and evaluated considers to desirable Bit Error Rate (BER) < 0.025, Signal-to-Noise Ratio (SNR) > 21 dB, a great amount of data payload that is 525 bit per second, and the noise caused by watermark is inaudible. All of the experiments, trials, and results show that the designed audio watermarking system is robust and imperceptible.

Lots of important, confidential and personal information are stored on the computer and sent on the internet need to be protected. Cryptography is one of technique to protect the file. This research proposes a cryptography method by combining the RSA algorithm and the RC4 algorithm. The power of the RSA method dependent on the key quality, which is obtained from the computation of two different prime number of p and q. The encryption result can encrypt the file byte but beyond the limit of the byte. RSA algorithm needs to be modified, so the key is more various and works on file byte. Encrypt file from RSA algorithm, combined with RC4 algorithm to get the higher security of file. To test the performance of the proposed algorithm, entropy measuring, avalanche effect counting, and bit error ratio (BER) measuring, and time required for the computation process were performed. Based on the proposed method, proved that file security has improved because the testing method produces a satisfying result.

Indonesia has held simultaneous general elections to elect the President/Vice President and legislative members on April 17, 2019. However, there were at least 4 (four) important problems arised in this election i.e. the problem of logistics distribution, the duration of the ballot counting that is too long, the inconsistent regulation of vote counting, and the error in votes recapitulation. Blockchain technology can be a solution to deal with those problems. In this paper, we propose a design of blockchain-based electronic voting system using Hyperledger for Indonesia case. Our design was analyzed in 3 (three) stages of analysis i.e. blockchain necessity, problem solution, and secure election requirements. Based on the results of the analysis, our proposed design can optimally apply the blockchain to solve the problems of Indonesia general voting system and can meet the secure election requirements to increase the trust of all the involved participants.

Cryptojacking (also called malicious cryptocurrency mining or cryptomining) is a new threat model using CPU resources covertly "mining" a cryptocurrency in the browser. The impact is a surge in CPU Usage and slows the system performance. In this research, in-browser-mitigation has been built using an extension in Google Chrome against cryptojacking using the Taint Analysis method. The method used in this research is attack modeling with abuse case using the Man-In-The-Middle (MITM) attack as a testing for mitigation. The design of the proposed model can notify the user if a cryptojacking attack occurs. That way the user can find out the script characteristics that run on the website background. The results of this research can mitigate cryptojacking attacks, from 100 random sample websites the proposed model can detect 19 websites indicated cryptojacking

The amount of data is experiencing rapid growth in this era. Data can be in the form of text, images, sound or video. Social media has become one of the factors of data growth. Every person in the opinion of opinion and complaining on social media from these opinions can be analyzed. In this study sentiment analysis uses the support vector machine algorithm. The first step is crawling data using the Twitter API with keywords. After collecting data, the preprocessing process is carried out, after the preprocessing process the feature is retrieved for each tweet, the features obtained are then collected into a feature list. The feature list is then transformed into a feature vector in binary form and then transformed using the Tf-idf method. The dataset consists of 2 data, namely training and testing. The training is labeled manually. To test the performance of the algorithm used the K-Fold Cross Validation method. The test results are obtained an average accuracy of 98% with the composition of training data and testing data. From these results the Support Vector Machine method can be used for sentiment classification of JNE twitter data.

Reliable robot communications is important for robot teamwork. Reliable communications is one of the main key which can be provided by multiple-input multiple-output (MIMO) technique. In MIMO antenna design, correlation is the main concern of MIMO antenna since each antenna should work independently. This paper proposed array antenna at 5.6 GHz with 4 planar dipole elements with very low correlation. These antennas are arranged on azimuth plane. We conduct computer simulation of antenna design and simple MIMO communications system. It is confirmed by simulations of communications system considering correlation between antennas that the proposed antenna has very low correlation and very close performance compared to ideal MIMO antenna.

Natural disasters are events that are difficult to avoid. There are several ways of reducing the risks of natural disasters. One of them is implementing disaster reduction programs. There are already several developed countries that apply the concept of disaster reduction. In addition to disaster reduction programs, there are several ways to predict or reducing the risks using artificial intelligence technology. One of them is big data, machine learning, and deep learning. By utilizing this method at the moment, it facilitates tasks in visualizing, analyzing, and predicting natural disaster. This research will focus on conducting a review process and understanding the purpose of machine learning and big data in the area of disaster management and natural disaster. The result of this paper is providing insight and the use of big data, machine learning, and deep learning in 6 disaster management area. This 6-disaster management area include early warning damage, damage assessment, monitoring and detection, forecasting and predicting, and post-disaster coordination, and response, and long-term risk assessment and reduction.

Autism Spectrum Disorder (ASD) is a brain development disorder that affects the ability to communicate and interact socially. There have been many studies using machine learning methods to classify autism including support vector machines, decision trees, naïve Bayes, random forests, logistic regression, K-nearest Neighbors and others. In this study provides a review on autism spectrum disorder by using a machine learning algorithm that is supervised learning. The initial study of the article was collected from a website provided articles were in according with this study, after going through the process of selecting articles 11 articles were eligible in this study. Based on the results obtained, that the most widely used algorithm in the literature study in this study is support vector machine (SVM) of 63.63%, with the application of machine learning in the case of ASD expected to be able to accelerate and improve accuracy in determining a diagnosis.

Every year, the number of cooperatives in the province of East Nusa Tenggara continues to grow. Cooperatives are present with the aim of helping the community on the financial side. The cooperative offers the principle of saving and providing low interest loans to its members. But there are times when lending is subjective. This condition is a major factor in the occurrence of errors in providing credit that leads to congestion (non-performing loan). This study focuses on the comparison of 5 classification methods using historical loan application data for a Multipurpose Cooperative in East Nusa Tenggara. The 5 methods are Naïve Bayes, K-Nearest Neighbor (KNN), Support Vector Machine (SVM) Random Forest, and C4.5. In the test results, it turns out that the C4.5 Method has better accuracy and a smaller error rate.

Real world imbalance data still become challenging problem in classifications. That is because of its imbalance ratio between majority and minority class and data distribution complexity. One of the solutions to deal with those cases is by modifying data using undersampling, oversampling or a combination of under and oversampling approaches. Randshuff (Random Shuffle Oversampling Techniques for Qualitative Data) is one of an oversampling algorithm which appropriate for nominal attributes. Randshuff uses IVDM (Interpolated Value Difference Metric) distance calculation and crossover with random shuffle technique. The problem arises where majority data contain distribution complexity problems such as small disjuncts, overlap and noise. So, evaluation of undersampling effects on Randshuff needs to be conducted. Several state of the art undersampling algorithms such as Tomeks links, Edited Nearest neighbors (ENN), Random Undersampling (RUS) and Near Miss which are combined with Randshuff were evaluated on five public data sets. The results show that RUS and Near Miss improve recall, f-measure and g-mean performance on Randshuff algorithm

Web phishing is one of the many crimes that occur in cyberspace and often threatens internet users around the world. Web phishing works by tricking the victim into a website page that has been designed to resemble the original page and then directing the target to submit the important information they have. Web phishing detection system needs to be developed to minimize attacks and theft of information using the website. Research related to web phishing detection system has been carried out by many researchers, one of them using data mining techniques, but still uses a single classification algorithm. Therefore, the addition of meta-algorithm is proposed to support the improvement of classification performance for the development of various web phishing detection systems. From the testing phase that conducted using Web Phishing dataset from UCI Machine Learning Repository, an increase in accuracy value of 97.1% is obtained by the addition of the bagging process, 97.3% by using the boosting process, and 97.5% by the addition of the stacking process. With the resulting improved performance, it is hoped that the model can be used as a reference in perfecting the development of various phishing web detection systems

There is a problem in the reduction of paddy fields and the effects caused by the reduced nutrient of paddy soils. There needs to be taken to maintain the sustainability of agricultural production. Strategic efforts are needed in the management of planting land with the aim that it can be utilized for optimal development of agricultural crops. One way that can be done is to arrange the planting of paddy and crops in turn. In order to obtain optimal planting results, it is necessary to select land that is in accordance with the criteria for growing crops to be planted. In this research, a decision support system prototype was developed to help farmers or users in determining the planting land that was in accordance with the criteria for growing crops. Using the topsis method and the addition of an expert system to the decision support system, resulting in a choice of land alternatives in the ranking system. From the testing of functional, the correctness of process and the user acceptance level, it is known that the decision support system prototype work process for the land selection of crops have gone well according to the criteria for growing crops to be planted in the form of land ranking.

Pornography is a global problem. Various efforts to overcome the spread of pornographic videos have been carried out by many researchers in the world with various methods such as the GLCM method for extracting image features based on skin color, the BossaNova method for extracting features in the images, combining two color spaces namely RGB and YCbCr and using segmentation skin color with YCbCr and KNN. Feature extraction used in the previous ones such as GLCM, YCbCr, RGB was vulnerable to changes in image size, translation, and rotation. In contrast to the SIFT and SURF methods that were used in this study. SIFT and SURF method invariant to changes in translation, size, two-dimensional rotation, illumination, can extract many key points and able to show object decryptors better. The number of datasets used was 7997 pornographic and non-pornographic images taken from Yahoo's dataset (NSW dataset) while 984 images were used as data testing. The feature extraction used the SIFT and SURF methods to get the keypoint and descriptor values. Descriptors of each image were clustered by K-Means. The center of each cluster was used as a visual dictionary's vocabularies. Then the image representation was performed using BoVW and finally, the classification was done by the KNN algorithm. For the results, the SURF method had better accuracy and time values compared to the SIFT method. The highest accuracy value of the SURF method was in dictionary size 300 with an accuracy of 0.8226 (82.26%). While the lowest accuracy value was in dictionary size 2600 with an accuracy of 0.6917 (69.17%). For the time variable, the SURF method was slightly faster than the SIFT method with an average value of respectively 2.49 and 2.52

Coronary Heart Disease (CHD) is one of the diseases that is the highest cause of death in various countries, including Indonesia. There are a number of factors that play an important role in the incidence of CHD, called the CHD risk factor. The technique that can be used to perform early diagnosis in identifying a person at risk of being exposed to CHD is classification. The research aims to implement the method of the synthetic neural network Self-Organizing Map (SOM) for the classification of CHD. This research is conducted using experimental methods, with stages including (1) data collection, (2) data pre-processing with reducing of attribute and normalization, (3) classification process with SOM method, (4) validation by holdout method, and (5) evaluation of the SOM method using confusion matrix. Evaluation is measured by calculating the value of accuracy, precision, recall, and error rate. The results showed that the most optimal level of accuracy in the data comparison of testing and training is 20%:80% with learning rate parameters 0.05, minimum learning rate 0.01, and maximum iteration 100. The values of performance measurement obtained are accuracy 62.5%, precision 60.33%, recall 63.33% and error rate 37.5%.

Class imbalances in data mining research are quite detrimental because there are difficulties in classifying minority classes (small number of instances) correctly. Oversampling is a method of balancing class distribution by randomly replicating instances in minority classes. multiclass classification might achieve a lower performance than binary classification as the boundaries among the classes may overlap. This issue may become more complex when facing imbalanced data. This study presents test results giving different k values in nearest neighbor searches which are used to balance the class using the ADASYN-N method. The results show that giving different k values can affect the performance of the ADASYN-N method depending on the number of dataset instances.

Social media is a reference material for companies to see customer behavior today. Analysis of data tweet can also be used as a reference to determine the business steps of companies in determining their policies. Twitter is one of social media that is often used. Users use Twitter to deliver their tweets to the general public. Airlines need feedback from customers to find out their views on airline services. The opinions or opinions of customers and even the general public are usually found in some social media comments. One social media that contains this opinion is on Twitter. This study aims to discuss aspects of the approach in the analysis of flight dataset sentiments. In addition, the feature selection process is also performed using the Mutual Information method.

SVM method can be utilized to classify opinion data based on assessment attributes to discriminate whether an opinion is classified as positive or negative sentiment. The advantages of SVM implementation such as generalization, its stability, classification, and ability to process both linear and non-linear data have made SVM being considered as a reliable classification method. According to these considerations, the authors carried out a sentiment analysis using SVM algorithm to evaluate Go-Jek service reviews. This sentiment analysis is expected to provide benefits for the stakeholders, particularly for Go-Jek. Support Vector Machine (SVM) algorithm is capable of overcoming high dimensional data set, classification problems, and both kernel linear and non-linear regression, which make SVM reliable for an algorithm for classification and regression. However, support vector machine has a drawback in selecting relevant parameters to enhance its optimization. A genetic algorithm method is necessary to resolve the problem of selecting relevant parameters in support machine vector method.

Along with the development of the times, the use of social media, especially Twitter, is increasingly being used. Of course this makes more people communicate on social media. Due to communication, it is possible that there will be utterances of hate speech delivered to certain parties, especially now that Indonesia is approaching the presidential election in 2019. The number of certain parties spread hatred in social media especially on Twitter. Therefore, as technology develops, we create a system that can detect a tweet based on the search for hashtag on Twitter whether it is classified as hate speech or not using the LSTM method as a classifier. The result of this system is to provide a label in the form of "hate speech" or "non-hate speech" on every tweet that becomes an input on this system.

Spam filters have become an important tool for ISP(Internet service Provider) to address the growing spam in cyberspace. Spam filtering is a program used to detect spam, prevent it and enter users' mailboxes. With the rapid growth of Internet technology, email communication is becoming increasingly important in people's daily lives. In addition, the amount of spam contained in emails has increased considerably in recent years. In this study, the authors will compare the performance of the decision tree classification models using or not using Recursive Feature Elimination (RFE) in spam filtering. The results of this test will be compared with the results of level accuracy, recall, precision, and F1 Score using or not using the RFE method. where in the results of this level measurement, the model used can be a reference or solution to filter spam correctly, especially through the use of supervised learning data.

This paper discusses the effect of different modulation schemes toward coverage area of an indoor Visible light communication. The idea of this research obtained based on the importance of using modulation techiques as needed. The main subject of this paper is power efficiency of each modulation refers to the coverage area. The modulation techniques used in this paper are On-Off keying Non return to zero (OOK-NRZ), On-Off Keying Return to Zero (OOK-RZ), Direct-current Biased Optical OFDM (DCO-OFDM), Unipolar OFDM (U-OFDM). This paper present the result of power efficiency, coverage area and Bit Error Rate performance of each modulation technique. The modulation technique that produces the best performance in this paper is U-OFDM. The U-OFDM modulation technique has the highest power efficiency and has the most widest coverage area.

the practical aspects of developing an Automatic Speech Recognition System (ASR) with HTK are reviewed. Steps are explained concerning hardware, software, libraries, applications and computer programs used. The common procedure to rapidly apply speech recognition system is summarized. The procedure is illustrated, to implement a speech based electrical switch in home automation for the Indonesian language. The main key of the procedure is to match the environment for training and testing using the training data recorded from the testing program, HVite. Often the silence detector of HTK is wrongly triggered by noises because the microphone is too sensitive. This problem is mitigated by simply scaling down the volume. In this sub-word phone-based speech recognition, noise is included in the training database and labelled particularly. Illustration of the procedure is applied to a home automation application. Electrical switches are controlled by Indonesian speech recognizer. The results show 100% command completion rate

Extracting aspects and opinion words is important in ABSA study. Generally, aspect extraction was conducted first then determining a pair of opinion words from the aspects. The process allows errors to occur. This study aims to obtain the appropriate pair of aspect and opinion words in opinion sentences using rule-based method generated from regular expressions. The strength of this process is that it enables to obtain a pair of aspects and opinion words simultaneously. This study is conducted in three stages, namely (1) selection of aspect-based opinion sentences in user reviews, (2) extraction of candidate aspects and opinion words in aspect-based opinion sentences and (3) aspect categorization. This study enables to extract pairs of aspects and opinion words using rule-based aspects generated from regular expression. Determining valid aspects uses aspect extraction. Evaluation of aspect categorization shows the value of precision of 0.82, recall of 0.70 and f-measure of 0.75.

PT PLN (Persero) Indonesia estimates the number of electricity customers in 2027 for the 82.11 million household sector and 7 million for the business sector. With the estimated increase, is there any influence on the amount of electricity distributed to each sector in 2027. So this study aims to see how much influence the number of customers on the amount of electricity distributed along with the prediction for the amount of electricity distributed in 2027 , carried out using linear and quadratic regression models. Then with equations using linear and quadratic regression models was obtained estimates of the amount of electricity distributed for the year 2027 with a linear regression model of 141.289,247 GWh for the household sector, and 106.974 GWh for the business sector. Whereas for the quadratic regression model 157.000,1 GWh for the household sector and 109.175,0 GWh for the business sector. To check the best model criteria from linear and quadratic regression models using the coefficient of determination. Based it, that value between linear and quadratic regression models, the most accurate results are quadratic models.

Over the past years, the popularity of mobile applications (or known as mobile apps) is continuously growing from year to year in the last decade. The development of mobile apps now becomes a part of business activity. It is essential for each developers to understand the users' need to keep their users on using their app. Sentiment analysis or opinion mining is an approach that can be used to analyze and conclude the opinion of the users of mobile apps. Recent method using lexicon based approach which is proposed by the previous study still has poor performance and still can be improved. There are two optimization chances that can be explored: lexicon resource usage evaluation and the application of domain specific features. This study tries to explore these two optimization chances in order to improve and optimize the method proposed by the previous study. The result shows that SentiWordNet can outperform other lexicon resources and the additional of domain specific features has a good impact on the performance of the classifier both on overall accuracy and f-measure average evaluation parameters compared to not only the results of this study but also compared to the result of the previous study.

Citation sentence is able to inform readers about relation between scientific articles that cite and are cited by finding its purpose against the research. Besides giving credit to other researchers and recommendation to read other related articles, citation can help readers to know what knowledge they have obtained based on the cited scientific articles they have read. In this research, we try to define citation categories for filtering scientific references which will be initial step in guided summarization of scientific articles. Our goal is to classify citation sentence first into 'problem', 'use' and 'other', then 'use' will be divided into 'useModel', 'useTool' and 'useData'. This category will make it easier to classify scientific articles into more specific topics. Then we use features namely voice, tenses, citation location, meta discourse and word vectors. Then, the testing of models show that the classification of citations using SVM Linear has the f-measure of about 67.7%. Moreover it is good enough when the test data is only one paper with the result of f-measure can reach 89.6%.

Part of Speech (POS) Tagging is a process of labelling word classes on sentences. One of the POS Tagging problems is some words that spelt the same but have a different POS Tag depending on the context of the sentence (ambiguity). The approach to solving this problem is using the Hidden Markov Model (HMM) Ngram Algorithm and the Viterbi Algorithm. This study discusses the development of a system for Indonesian POS Tagging using the HMM N-gram algorithm (Bigram and Trigram) and the Viterbi algorithm and compares the result between the HMM Bigram and HMM trigram. An Indonesian language corpus that has been manually labeled called Indonesian Manually Tagged Corpus is used as the knowledge for the system. Then the corpus is processed using the HMM N-gram algorithm to get the rules. Furthermore, process the data with Viterbi algorithm using the previous formed rules to determine the POS tag with the highest probability. The highest accuracy results is 77.56% using the HMM Bigram - Viterbi Algorithm. While the HMM Trigram - Viterbi algorithm has the highest accuracy of 61.67%. The result shows that the system can solve the problem of tag ambiguity with HMM Ngram - Viterbi algorithm and the accuracy of HMM Bigram is better than the HMM Trigram.

This research carried out the label aspect classification, sentiment analysis, and topic trends on the Open-Ended Question (OEQ) section for Student Feedback Questionnaire (SFQ). Multi-Class aspect label classification for SFQ will choose the best classification model by comparing the results of the evaluation of accuracy, precision, recall, and F1-score for each feature combination and comparison of four classification algorithms namely Decision Tree (DT), Naïve Bayes (NB), K-Nearest Neighbor (KNN), and Support Vector Machine (SVM). The results of this research are Classification Techniques using a combination of features of TFIDF, Unigram, and Bigram with the SVM algorithm which is the best Multi-Class classification model for labeling SFQ aspects. In addition, the SentiStrenghtID algorithm used to get sentiments and also the LDA (Latent Dirichlet Allocation) used to get annual topic trends on each survey aspect label. The findings can help Higher Education to support decision making in taking proactive actions towards improvement for self-evaluation and quality.

In the era of Industrial Revolution 4.0 marked by the use of information technology, the Internet in particular, university executives, as well as executives in companies, are often in such a situation which forces them to conduct business intelligence activities ranging from simply accessing data and information to analyzing them to support their decision making whenever and wherever they are. Typically, university executives are able to access data and information not only from their university databases through their web applications but also from other data sources using internet browser installed on their laptops or mobile devices. However, information access model using web application tends to be rigid where the order of menu, layout, and user-interface provided are static which forces the user to memorize or search for the right place to find the data and information needed. Certainly, this model cannot keep up with the dynamic information needs. This study produces a mobile-based application named m-BELA which functions like a human personal assistant equipped with pre-defined data, information, and knowledge who is able to interpret user's voice command and then responds appropriately by following conversation or providing data and information relevant to the command. This new solution offers a competitive advantage since the user does not need to load a browser, open the web application, and then search for the menu and UI which provide information needed.

The use of three phase induction motor as the electric drive is very popular because of the cheaper price, robust construction, and free maintenance, but it has a characteristic that is not linear so it is hard in the control. Some of the developed methods are expected to obtain better control. In this paper present of Indirect Field Oriented Control (IFOC) based on PID control used speed control of three phase induction motor. Using these controls can help get better output response with short time to reach the speed reference. The simulation result, Speed reference 700 Rpm needed 0.18 sec to achieve speed reference and 0.95 sec to steady speed. Then 800 Rpm 0.19 sec to achieve speed reference and 0.88 sec to steady speed, 900 Rpm 0.2 sec to achieve speed reference and 1 sec to steady speed. Design of IFOC+PID control a have better system response, evidenced by very short time to achieve speed reference and steady speed condition.

Keywords- Induction Motor, Speed Control, IFOC, PID.

In a wireless adhoc network, routing protocol plays an important role to find communication path. In this research, we investigated proactive and reactive routing protocol performance in a more realistic environment with dynamic network topology scenarios. We compared Dynamic Source Routing (DSR) as reactive routing protocol and a proactive routing protocol Optimized Link State Routing (OLSR). We used Rician and Rayleigh propagation model as a multipath fading model for more realistic environment in mobile adhoc network. We focus on network topology change scenarios by increasing number of nodes and behavior of the node. The results shows that OLSR have advantages which gain higher throughput and smaller end-to-end delay compare to DSR in Rician and Rayleigh environments due to multipoint replying (MPR) strategy. However DSR have more efficient in packet routing control with smaller normalized routing load. For dynamic topology change in Rician and Rayleigh fading, OLSR performs more stable and can gain higher performance than DSR.

Today, the growth in the number of users is increasing with the development of advanced technology. In addition, this is also accompanied by higher community needs due to technology that is able to support the needs of the community. IEEE 802.11ah is a standardization of the development of IEEE 802.11 that is able to overcome these problems, because the coverage area of the transmission range is up to 1 km, has lower energy consumption, and is capable of serving up to 8191 stations (STA). In this research, an analysis of changes in modulation and coding scheme (MCS) and restricted access window (RAW) Slot will be carried out on the IEEE 802.11ah standard for VoIP services by comparing two codecs, namely G.729 and G.711. In this research, we used network simulators NS-3 on VoIP services by comparing two codecs namely G.729 and G.711 and investigated how they affect network performance. We compare the average delay, throughput, and packet delay ratio (PDR) as network performance parameters.

In Indonesia was still many handling of natural disasters using traditional methods. Therefore increasing the ability to handle natural disasters through sharing experiences or research in this field was very important. Several studies had been conducted in the field of natural disaster detection system using internet network. But natural conditions and internet network blank spots became obstacles of the disaster detection system in remote areas. So it was needed a system with flexible network, energy efficient and accurate in data sending. This research propose a system of flooding detection with mesh network, ZigBee device and sensors of water conditions. By using mesh network and ZigBee take an advantage to be implemented in the remote area. Sensors would detect the indication of flooding and send the data trough ZigBee. The experiment showed that mesh network with the distance between nodes max 75 m was the best result and condition.

Filtering is a necessary part from the digital signal processing in radar. This research is focused on comparing the performance of 4 windowing methods for designing finite impulse response (FIR) digital filter for a 9,3 GHz surveillance radar system. In this study, Blackman, Bartlett, Hanning, and Kaiser window type are modelled in QT software. The experiment is done by taking 512 and 1024 frame length in all frequency bands. Parameters examined are the amplitude and the frequency. Result showed that Kaiser has the best performance for this radar application.

Thesis / scientific work is a science composition that presents facts and is written according to a good and correct writing methodology. But to find out what the theme discussed should read the abstract of the essay. Along with the development of information technology that offers convenience for human labor in simplification of his work. Difficulty in determining the theme in the thesis can be aided with applications that are able to set themes based on abstract. To make the application of thesis thematic determination can use string matching algorithm, one of string matching algorithm is Rabin-Karp algorithm by doing string matching based on the hash value on text and hash value in a pattern. The Rabin-Karp algorithm is often used to detect plagiarism text document, comparing the overall hash value of k-gram contained in all text documents, then deciding the degree of similarity. Then it can be concluded that Rabin-Karp is able to string a text as a whole, from this the researcher is interested in proving what if the Rabin-Karp algorithm is implemented in the classification field wherein in this study it is used to string matching each word or keyword refers to the theme of an Indonesian essay or thesis abstract. The application is built using web programming language PHP and MySQL database. The confusion matrix is used to determine accuracy, obtained 88% and 12% for misclassification rates

From the previous study related to web phishing, the researchers overlook the imbalanced class problem on the dataset. theoretically, the majority of classification methods would assume that the nature of the class distribution is balanced. It caused the classification's performance of the method will be declining. Therefore, the mechanism of imbalanced class handling is severely needed. In this study, OSS and SMOTE are used as the resampling method to handle the imbalanced class problem. Those algorithms work to balancing the class distribution of the dataset so that the accuracy and the g-mean score of the classification will be enhanced. Based on the result, the ensemble of those methods (OSS and SMOTE) can enhance the classification's result significantly either on binary type class and multiclass type dataset. Hence, the combination of OSS and SMOTE can be a plausible option to handle the imbalanced class problem on the web phishing classification either on binary class and multiclass datasets.

This paper proposes a hybrid Intelligent Prediction model, which optimizes the Triple Exponential Smoothing (TES) alpha, beta and gamma parameter algorithm using Fruit Fly Optimization Algorithm (FOA) algorithm to predict time series data. Based on the research of previous authors, the TES algorithm is very likely to be sensitive to changes in the constants to the 3 parameters, where to do benchmarking it results use the MAPE method. Therefore, the authors limit this research by optimizing the parameters of the TES algorithm with the FOA algorithm. The dataset used in this experimental study has datasets which obtained publicly from the data market website repository. From this study, it was found that combination of FruitFly Optimization Algorithm - Triple Exponential Smoothing (FOA-TES) can predict the time series data well with the average MAPE of 6%, better than the TES with an increased MAPE as 4%.

Waste management, especially waste collection issue, has been an ultimate global issue. Particularly, problem of collecting waste in minimum time, especially in big cities, becomes a challenge that needs solution. In this paper, this problem is approached by a well-known problem called Vehicle Routing Problem (VRP). The objective is to generate collision-free trajectories along the process of collection while minimizing the travelling time of the slowest vehicle. A novel method to solve such the problem, namely, Augmented-Node-based Ant Colony Optimization (ANACO), is introduced. The main feature of this method is the generation of trajectories, i.e., paths and velocities, of all waste-collection vehicles. The advantage of the proposed method is in the ability to provide velocity information besides path, which is totally different to the typical ACO algorithms that only provide path information. By the proposed method, the travelling time of the slowest vehicle can be minimized. Therefore, the requirement of fast-collecting capability can be satisfied. Simulation results are presented proving that a minimum-time and collision-free trajectories can be generated to solve the problem.

Mobile robots have been widely applied to various aspects of life. For example, a robot for detecting gas leaks. The process of finding this gas leak can be done by mobile robots automatically, but in the search process, there are three challenges to be faced. Namely vast search space, complexity in handling unstructured objects and a certain level of intelligence is needed, so that the search process on mobile robots to be optimal. To answer these challenges, control methods will be developed with the Fuzzy Hierarchical Model (FHM) and Particle Swarm Optimization (PSO). Where the methods are expected to be able to control mobile robots to search for gas leaks, in a complex specific virtual environment more optimally. In this research, the design of the FHM-PSO control model was carried out and then applied to an FHM-PSO software application that can control the swarm robot model until it successfully reaches the gas leak center in a specific complex virtual environment optimally.

The dropout student is a case that should be a concern at a college. The uncontrollable dropout will affect the quality of the universities. Dropout are done for a variety of reasons, one of which has been carrying out the maximum study period. The bachelor program have maximum study period must to be pursued is 8 years. In this research, predictions of student dropout is done to students who have the possibility of exceeding the maximum study period. The predictions are done by digging the data patterns on the student's academic database by leveraging the academic achievement index data of each semester and class attendance This research aims to accelerate training and improve the accuracy of prediction with backpropagation (BP) optimized with particle swarm (PSO). Further accuracy results will be compared with a simple backpropagation algorithm without optimization. The backpropagation algorithm, which is optimized with particle swarm optimization, generates the best accuracy of 92.85% with 112 epoch. Meanwhile, the backpropagation algorithm generates an accuracy rate of 84.28% with 1000 epoch. The highest accuracy obtained using the best architecture is 8-15-2 and the number of particles 50. Optimization of network weight backpropagation with particle swam can improve accuracy and the number of iterations achieved decreases

Electricity is an essential need for humans, specifically for the lighting function. Lighting is needed to aid humans visually perceive objects in areas that require lighting, both at night and during the day. At this day, there are two types of lamps that are extensively employed, particularly the type of Light-Emitting Diode (LED) and Compact Fluorescent Lamp (CFL). The adoption of lights in an area takes a lot of energy, so it is particularly crucial to choose the appropriate type of lamp that is adequate to perform effectively and efficiently. To determine the light distribution of the two types of lights, it is crucial to investigate the distribution of light. The investigation was carried out by the direct measurement technique on 6 and 8 watt LED lamps, as well as on CFL lamps 11 and 14 watt using luxmeter in an area projecting 3 x 4 meters. The analysis range is varied in several horizontal and vertical points. From the assessment results it was noticed that the LED and CFL lamps the value of the light intensity reduces if the horizontal and vertical distance of the measurement point are far from the lamp. The ends of the light distribution mapping discovered that the light distribution on the LED lights is better indeed and brighter at each point. This is because all the electrons recombine with holes. Whereas in the CFL lamp there has been a filament heating and ionization between the electrons with argon gas and mercury vapor which generates another energy which is relatively large heat energy.

A shortage of reactive power in the power system will caused voltage instability. In order to improve the voltage stability of power system, one of FACTS device which is STATCOM can be used to inject reactive power. This paper focused on the effect of injecting reactive power from STATCOM in power system for voltage stability and is tested on IEEE 9 bus test system. Fast Voltage Stability Index (FVSI) is used in this paper to identify the weakest bus of the system. As a result, the analysis of voltage stability and the subsequent installation of STATCOM are tested on the bus-5. The P-V curve method is used in analyzing the voltage stability of the power system. The result shows that after the installation of 80 MVAR STATCOM to the system with a 250 MW load improves the voltage profile from 0.86 p.u. to 0.91 p.u. in normal condition. And in contingency, the system is secure after STATCOM installation. The capability of bus-5 also improves both in normal condition and contingency with the installation. However, the installation of STATCOM at bus-7 did not affect the voltage profile of the system. Therefore, the location of the STATCOM installation needs to be considered.

Nowadays, everyone needs a computer or PC to help them run daily errands or work. Any device on a PC has its own performance level, and can be maximized through a method called overclocking. Memory or commonly known as RAM is the easiest device to overclock. By using Memory Try-It feature in BIOS, user can obtain a better performance compared to its default condition. The authors used RAM Galax Hall Of Fame 16 GB. HWBOT Realbench benchmark was performed to observe the increased performance of RAM during overclock condition. Benchmark test results a system score which can be utilized as a reference for the user to understand the increased performance of a device (in percentage) after being overclocked.

This research was conducted to analyze the Review and Rating of Jakarta's consumer confidence in conducting Queue Online Booking based on the Technology Acceptance Model (TAM). The variables used are Perceived Ease of Used, Perceived of Usefulness and Behavioral Intention to Use of the main Technology Acceptance Model (TAM) variables that can be influenced by Trust, Review and Rating. This study aims to determine the impact of the Review and Rating by adding a new variable that is trust. Respondents in this study used the Structural Equation Modeling (SEM) method with 110 respondents collected from internet users who have made online bookings. Respondent data is reduced because some do not comply with the provisions of the study to 100 respondents. The results obtained from this study proved that trust variables influence and increase consumer confidence in conducting online booking queues. Therefore, online marketplace companies must make and improve reviews and ratings as one of the main marketing tools to increase consumer confidence and company revenue.

In 2018, e-commerce activities contributed USD 12 billion to Indonesia's digital economy activities. On the other hand, some internet users stated having problems with transactions and mistrust with e-commerce. Tokopedia is one of the largest e-commerce in Indonesia that has the largest monthly web visitors and dominates the ranking in Google PlayStore and Apple App Store. This research aims to analyze the effect of security and trust on buying decision on the Tokopedia mobile apps in Jakarta. The research was conducted using quantitative method. Research data was collected by distributing online questionnaires through Google Form in August 2019. The research model was analyzed using PLS-SEM of 129 respondents. Results show Trust and Ease of Use have significant influence on Intention to Buy, and Intention to Buy has significant influence on Buying Decision.

Information Technology learning content on Youtube is huge, both produced by individuals and companies, from professionals to academics, this greatly enriches the learning resources specifically in the field of information technology, from the many contents, what kind of content can be accepted by students majoring in information technology as a learning material which will ultimately increase the view of the video content and help students in their academics. This research was conducted to examine the factors that influence learning interest through Youtube media on Information Technology students at Bina Nusantara University. In addition, this research is to ascertain what factors influence directly or indirectly through YouTube media learning. This study uses a TAM model that aims to determine interest in using Youtube as a learning resource. Data samples were collected from 100 respondents of the Department of Information Technology at Bina Nusantara University. The results obtained from the installation of structural equation models in the sample indicate that behavioral intention is significantly influenced by attitude to use Youtube as a Learning Media for Information Technology. The overall findings show that students start with a motivation, then develop an attitude to use youtube, and driven by that, it's then go to intention to use youtube as a learning media for Information Technology.

Event Organizer is a professional service provider of organizing events in order to organize the entire series of events to suit the expected goals. This study aims to determine the Event Organizer integrated with e-commerce such as Tokopedia, Bukalapak, Blibli, Lazada, Shopee so that it can be accessed by customers and vendors providing event needs easily. Some factors that influence customer behavior so that they are interested in using Event Organizer services are by using Perceived Value, Service Quality, User Satisfaction, Trust, and Purchase Intention variables. This research produces several hypotheses that will illustrate the interrelationship between variables and the effect on variables. The hypothesis will be validated using an online survey involving 100 respondents. The results of the questionaire will be tested using SmartPLS software so that it is accurate. This study shows that many factors influence customers in deciding to use Event Organizer services. The results showed that the influence of trust has a positive effect on purchase intention.

Telegram is a cloud-based instant messaging application that focuses on speed and security. Telegram is designed to make it easy for users to send text messages, audio, video, images and stickers to each other safely. Telegram is growing rapidly and the application has been downloaded more than 100 million times. Although it has been downloaded more than 100 million times in reviews on the Google Play Store some users complain about the Telegram application. The problem is part of the usability heuristic interface aesthetic and minimalist design problem that exists in telegram applications. Because of this, it is necessary to explore more comprehensive problems to better explore the problems that occur in telegram application users. The usability heuristic level needs to be known to explain its effectiveness in telegram applications for users. This study provides evaluation results to determine the level of usability heuristic in telegram applications using multiple linear regression methods. The results obtained from this study indicate that 5 usability heuristic variables exist in telegram applications according to the user.

This paper proposes a novel research proposal to explore users switching platforms behaviour within the context of social network sites. The aim is to investigate the key factors essential to social network users' decision to switch platform. In this regard, a mixed methods approach of an exploratory - qualitative study (interviews), followed by a confirmatory - quantitative study (survey) will be proposed, hence so called a mix-method research design. The proposed study also specifically aims to contribute to Information Systems literature by adding value to the users' switching decision in social network sites platform, and extending the use of switching cost theory. In particular, by anticipating the salient factors affecting the social network sites users to switch platforms, as well as interrelationships using a mixed method approach.

Detecting fraud in fingerprint registration poses a unique challenge as we cannot rely on an existing employee's attribute. Furthermore, analyzing using supervised algorithm cannot handle unlabeled data that generated uniquely for this case. We study the patterns of employee's presence and activity report data and found that fraud action tends to be closely similar to other fraud action. Therefore, we propose a tight clustering method to detect fraud in fingerprint data using DBSCAN algorithm, as tight distance calculation removes non-fraud data because non-fraud data is generated to be unique naturally

The development of game technology is currently running very rapidly, especially mobile-based games. Such large Android smartphone users influence the direction of Android-based game development. Google Play Store is a provider of applications to games both paid and free for Android users. Application development companies competing to take advantage of the marketplace to benefit. To get sympathetic from the user, the developer companies do various ways both from the marketing side to the quality of the gameplay and design in order to become a popular product, especially on the Google Play Store. Research conducted at the moment is to analyze the popularity of games of each genre on the Google Play Store based on the dominant game elements to prove that whether the dominant game elements affect the motivation of players so that the game reaches the standard of popularity. The Sillaots parameter is used to analyze the dominant game elements and then the results are analyzed using the Octalysis framework that is tested on gamers without knowing the results of the previous test. The results of this study reveal that the relationship between dominant game elements and motivation reaches a value of 100% that it is true that each genre has different characteristics of each genre and motivation is different both in terms of design and game performance.

Pregnancy monitoring is important to ensure the sustainability of a nation. Local healthcare (PUSKESMAS) responsible for pregnancy health monitoring and maintenance in certain region. Pregnant woman activities including walking has long been observed to have positive correlation with the pregnancy health. The advance development of smartphone enables the development of human activity in real time based. The activity can be reported and plot on the top of geographical map and provide rich information to improve the quality of pregnancy health managements. In this paper we present our solution to the pregnancy mapping in real time online monitoring of their activity. In order to ensure the alignment between business need and the stake holder expectation an assessment of the system is carried out. User Experience is important success factor of system implementation. User experience questionnaire (UEQ) is a generic tool to measure user experience. In this paper we present our online real time mapping and monitoring system and user opinion about the implemented system. According to the questionnaire result, our proposed application considered to be in moderate level but still need improvements in many aspects.

There are still a lot of excessive use of lamps, televisions, air conditioners (AC) and other electronic goods, resulting in a surge in electricity bills charged by electricity users due to neglect and waste of electrical energy. Based on these problems a system that is needed not only to monitor but also to control electrical equipment remotely so that electricity consumption can be controlled.

Wireless Sensor and Actuator Network (WSAN) technology is able to monitor the physical condition of the environment which is widely applied in intelligent environments. WSAN is placed at certain regional points that will be observed the physical condition of the environment, each WSAN can use several sensors and actuators which will later be sent to the server via a wireless connection.

In this research, we will test by making Smart AC (Air Conditioner) where at every point where there is AC will be installed WSAN. Data from several sensors generated from WSAN will be sent to the server to be observed and processed using intelligent computing and machine learning (K-Means and Naïve Bayes) so that the AC can turn on and off according to the physical conditions in the place.

Transmission Expansion Planning (TEP) is a plan to add an electricity transmission network that aims to minimize the investment costs of network development by taking into account the power losses on the transmission line. TEP consider to and fulfil the requirements such as technical, economic, and reliability of electric power systems. In this study, the Genetic Algorithm (GA) method is purposed to minimize investment costs taking into account power losses in the Garver's 6-Bus transmission network system. The optimization method using GA is selected since it provides solutions to non-convex problems found in TEP. In this study, the simulation on the Garver's 6-Bus network system namely a voltage level of 400 kV. Furthermore, future values factor are used in this study. The simulation result in the Fitness cost of the transmission network at a voltage of 400 kV, the Fitness cost is 1.849x109 US$ with a 1.592x106 US$ of a deviation standard and 1.552x106 US$ of RMSE.

Generally, photovoltaic were connected to the Maximum Power Point Tracker (MPPT) Controller to increase and maximize the output power of solar panels. But we know, that the output power of photovoltaic will be influenced by the amount of irradiance captured on the surface of the photovoltaic. The accurate measurement of irradiance value is an important point for evaluating and developing photovoltaic energy systems. This research shows how to estimate irradiance based on the maximum power point current of photovoltaic on the MPPT controller in real-time. This irradiance estimation is based on mathematical modeling of solar panels. This research shows that the estimated solar irradiation produced by measuring the maximum power point current has an average error 1.68% to the actual solar irradiation.

Fierce competition on ride hailing applications in Indonesia, encourages the development of the application that are not only reliable but also have to be customer-oriented. In the context of gig economy, customers consist of clients and gig workers. Lack of customer engagement can reduce competitiveness, images, and profit. In 2017, Indonesian Consumers Foundation exposed that 13 percent ride haling customer's disappointment was caused by the application. This research evaluated Electronic Customer Relationship Management (E-CRM) features in ride hailing applications, then finding out whether those effect electronic service quality (E-SERVQUAL), in another words, customers perception. Data analysis is done using Partial Least Square - Structural Equation Modelling (PLS-SEM) method with total number of valid responses processed was 204 respondents. The results implied that navigation, privacy and security, online community, and customer service are the E-CRM features that really need by the customers. By knowing the result, the researchers want to show the asymmetry signal between the providers and customers perception. Thus, it can lead providers to develop better E-CRM features in ride hailing application, conquering the customers dissatisfactions, and becoming basis theory for connecting the concepts of loyalty, engagement, and gamification to clients and gig workers in the context of gig economy.

This paper present Point Cloud maps using ORB-SLAM with visual sensor or monocular camera like the webcam, including map reuse, loop closing and relocalization capabilities. The system integrated with ROS works in real-time on raspberry pi based mobile robot from small hand-held indoor sequences. This system performs relocalization and loop closing with a high invariance to viewpoint in real-time. For large scale is addressed building a covisibility graph that allows the systems to perform most mapping and tracking operations locally. The ORB SLAM method is used and it detects targets which will help the mobile robot to explore unknown environments. With this exploration process, mobile robots will navigate autonomically from the initial position to the location to be mapped. The entire design of this mobile robot realization will be designed using a framework that is the Robot Operating System (ROS) with the help of tools that are already available in it. For results in the realization of SLAM visuals using the ORB SLAM method can detect objects optimally with an object size of 31.5x56 cm at a distance of 70 cm and get an error value of 1.21% and an accuracy value of 98.79%, and the system can work in unknown environments and get a visualization map that comes from the camera sensor.

Blockchain the new technology that rapidly develop contemporary beyond the cryptocurrency. Several researches reported the adoption of Blockchain Technology for health, pharmacy, logistic, and education sector. However, the report research in education sector that adopt Blockchain Technology for pharmacy very rare, therefore this research tries to explore on how the Blockchain technology influence the four critical success factors of pharmacy industry which are people, process, technology and organization factors or commonly known as Leavitt diamond model. The systematic literature review used in this research and found a thousand and two hundred ten (1210) papers from eight (8) reputable publishers and an index such as Taylor and Francis, IEEExplore Online Library, MDPI, ACM, Science Direct, Wiley Online, Emerald Inside and Scopus as Index paper publication. Eighteen (18) papers were selected, 23 critical factors found that will influence to people, process, technology, and organization.

Data Center Network (DCN) is a popular approach to build a big and scalable network, which processed big data. One of the common topology used for DCN is Fat Tree, it is consists of a core switch, aggregation switch, and edge switch interconnecting with each other in a hierarchical manner. Each DCN topology includes Fat Tree, uses routing topology to make data transmission between each node in the network. DCN is a big network and so it is a good way to simulating some topology scheme in a simulator before implementing it to the real world. NS3 Simulator is the most widely used Network Simulator, well known for its discrete-event flexibility. In the paper, we present a framework for Fat Tree topology with XPath routing protocol in Network Simulator 3 (NS3). The information we provide includes performance evaluation of XPath routing protocol in Fat Tree DCN and tools available in NS3 to study routing protocol performance

A Human resource in a company is a person in charge of finding new workers. Get a qualified new workforce, a human resource must be right in selecting new job candidates in terms of ability and personality. This study provides an alternative perspective for a human resource in getting one's personality data through their tweets on a Twitter account. This study uses the Naive Bayes Classifier algorithm with W-IDF (Weighted-Inverse Document Frequency) weighting to classify the personality of new recruits into one of DISC's personality theories, namely Dominance, Influence, Steadiness, and Compliance. By using training data and test data as many as 120 personal Twitter accounts and labeling of words that have been verified by psychologists, obtained personality distribution. The classification of the tweet data is Dominance 90 accounts, Influence 10 accounts, Steadiness 8 accounts, and Compliance 12 accounts. Evaluation of the accuracy level of 36.67%.

Diabetic retinopathy is one of complications diabetes mellitus that most common cause of permanent blindness. Diagnose of diabetic retinopathy conventionally is considered less effective than utilizes computer-based system as an analysis techniques used data mining as alternate process for diagnosis. In this research aims to know the level of accuracy and computational time. This research used combination of artificial neural network algorithms and correlations-based features selection (CFS) on the messidor-features dataset. The result showed the accuracy without feature selection about 72.02% with running time of 2.46 second and applied feature selection the accuracy was 73.24% with running time 0.98 second. The evaluation result using ROC curve shows that the combination of artificial neural network algorithms and CFS categorized as fair classification.

Watermarking is a method to protect the copyright of a work. Watermark can be either text, image, audio, or video. Data insertion is done in such a way so that the data will not damage the secured digital information. The data that is inserted cannot be removed from the digital information and must be able to be re-extracted. This research will analyze the compressive sampling based on DCT-DWT for watermark compression on video watermarking using DWT-SVD and OMP reconstruction with RS Code to check on the error bit. The result will be video quality with 0.25 of BER, 54.577 dB of PSNR, and 0.215 of MSE. The system will withstand the rescaling attack.

Trending Topic is one of the features found of Twitter in a short text. However, the short text used as a trending topic on Twitter sometimes confuses its users, so they need to be classified into several labels, but one tweet can have more than one label called multilable. The lables are politics, sports, entertainment, tourism, business, and other news. Another problem is the multi-labeling of classifications. single- label will classify a trending topic into one label, while multi-label classifies into more than one label. This paper aimed to classify Twitter's trending topic using Multinomial Naive Bayes (MNB) for single-label data and K-Nearest Neighbor (KNN) for multi-label data. The steps were to collect trending topic data along with their tweets, labeling and text preprocessing, weighting TF-IDF, single-label classification using MNB and multi-label classification using KNN with the Binary Relevance approach, finally evaluation and analysis of results. By using K=3, the results show that KNN have 88.05% accuracy for multi-label data, whilst, MNB has a good result for single-label data 82.53% accuracy.

The transport industry is Indonesia's largest consumer of fossil energy. Electric vehicle (EV) is an attempt to decrease the use of fossil fuels.Types of electric motor often used is an induction motor. In this paper, the parameter design and performance of squirrel cage induction motor 5 HP for EVs are investigated by changing the voltage system, 48 V for low voltage system and 360 V for high voltage system. The design process is performed using simulation software. With the same design factor, main dimensions of two voltage system designs have same size. The Different design located on stator winding and slot. In stator winding, wire gauge diameter and number conductors per slot are different. In stator slot, low voltage system needs a wider slot than the higher voltage system. The performance from ANSYS Maxwell simulation of high-voltage and low-voltage induction motor three-phase frequency 50 Hz output 5 HP for golf cart electric vehicles voltage system does not have a significant difference. The two design have efficiency, power factor, rated torque and slip about 87%, 0.93, 24.5 Nm and 3%.