Clustering Dependent Recommender Systems

Mohammed Abdul Hameed (Osmania University, India)

A number of clustering dependant recommender systems (CD_RS) have been consider in this paper including IGCRGA (Information Gain Clustering through Rank Based Genetic Algorithm), IGCEGA (Information Gain Clustering through Elitist Genetic Algorithm), IGCGA (Information Gain Clustering through Genetic Algorithm), IGCN (Information Gain Clustering Neighbor), among others. Non clustering heuristics used in recommender system - namely, entropy and popularity - were also considered for comparison purposes.

The two evaluation metric, Mean Absolute error (MAE) and Expected Unity (EU) used in this work show that, firstly, CD_RS out perform their counterparts which used Non clustering heuristics; secondly, IGCRGA emerged vector amongst the CD_RS.

Journal: International Journal of Simulation- Systems, Science and Technology- IJSSST V13

Published: Feb 27, 2012

DOI: 10.5013/IJSSST.a.13.01.01