Similar Cluster recommendation of Product Purchases by Pages liked Analysis

Abstract

In recent years, social network become more popular and influential on the online commerce. Many customers can find the detail of products and trends on the social network using it to make the decision whether to purchase the product or not. In addition, it is also commonly used to collect users' related data to predict their future actions. In this paper, we study the correlation of purchasing histories and the customer's interests on social network (Facebook) to predict the future interests of customers that would show through social network. In addition, we use the data to find more similar target group of customers who are also interested in the same category of product by text analysis with Facebook Pages Liked. We focus on the product categories aligned by the shopping online websites and find the customer cluster which is similar with the customer who purchased. The study results are beneficial to focus groups of customers for communication or suggestion about the products matching with the groups of customers who are interested in the same group of products and have high potential to buy them. This model can adept with many real-world businesses to identify the user's characteristics and find more similar group of customers who interest in the products by changing the user attribute to find the user's interest. In our study, we adapt this model with the prototype system to recommend the similar group of customers.