A HYBRID PRODUCT RECOMMENDER MODEL FOR BUSINESS APPLICATIONS

Today customers are highly exposed to the essential amount of information related to products and services like never before. This leads to an enormous amount of diversity leading to consumer's demand, thus becoming a challenge for the retailers to cater right products and services as per customer's preferences. Customer reviews, opinions and shared experiences related to a product turns to be an effective source of information about customer's preferences that can be utilized by recommender systems. To recommend products to the users, buying list of users, viewer's list and purchase count of the products are considered as main fundamental attributes for conducting the analysis of the products purchased and viewed. In this paper, a hybrid recommendation model that combines machine learning, collaborative filtering and data analytic is proposed. The recommendation algorithm begins to acquire cluster of similar users. Further, Similar shopping basket of customers is prepared with User-item matrix and purchase count of the products. The data set is then clustered as per the requirements to form train-test data. The Experimental Results shows low error i.e. lower root mean square and has high precision and high response time in comparison to Popularity based (Baseline) recommendation model