期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2018
卷号：18
期号：2
页码：91-102
出版社：International Journal of Computer Science and Network Security
摘要：Mobile phones have become vital in social life. The obvious market competition complemented by advanced functionality and feature is building consumers’ mobile phone decision making difficult and challenging. In this study, a user preference /voting based methodology was applied to build a recommendation system for mobile phones selection. A model with user preferences and actual weights of different variants of mobile phones was exercised to group the mobile users according to their interests and to cluster the mobile types based on the weights obtained from the actual attributes linked with user voting. The initial selection of cluster centers is a potential limitation of the popular k-means algorithm. In this paper a variant of clustering process is proposed with better initialization of cluster centers as representatives. The recommendation system was evolved by conducting a controlled experiment that involved 1000 mobile phone users. The information was gathered online through a well organized questionnaire. The analysis results indicate that the use of the proposed system results in higher satisfaction than equal-weight based benchmark systems.
