A Framework for Analysis of Purchase Dissonance in Recommender System Using Association Rule Mining

Recommender System plays a vital role in various domains like E-Commerce, Entertainment, and News and currently in social networking. It is used as a tool in marketing for attracting the interest of customer. The purchase patterns of customer dependent on both psychological and external attributes. Generally customers' purchases vary with respect to their mood swings. It becomes a need for Recommender System to incorporate psychological concepts for understanding customer role for predicting user interests. The objective of the paper is to design a framework for Recommender System incorporating cognitive dissonance a psychological factor. The recommender system is designed as a hybrid system which combines both content and collaborative concepts using association rule mining concept a data mining technique. The experimental result of the framework are tested and analyzed for mobile industry to analyse the dissonance in choosing mobile tariff plans. The experimental results it is found that customers are affected with 16% of post purchase dissonance in opting for mobile tariffs. It is found that the recommender System when designed with psychological inputs provides a better mechanism for making effective decision for introducing new marketing strategy.