PurTreeClust: A Clustering Algorithm for Customer Segmentation from Massive Customer Transaction Data
The Main aim of this Project is to provide the Recommendation to the user based on User Preference and User life Style. And to Cluster the Customer Group based on the Transaction data processed by the Customer.
Identify the Customer POI in retail and E-Commerce companies is became a challenging for Companies, So most of the E-Commerce Companies will asked the user POI when they login or Recommended the Product that is not in the wish list of the particular customer. And the Product that Companies will Recommended would be based on the number of positive rating to any particular product, so that many product in the E-Commerce website would became a Cold Items (The Product that are not Sold for quite Long Duration). Companies will not recommended any cold rated product to any user until the number of rating to the particular product would increase that would make the cold rated product will be in stock for long time, and moreover the product that are recommended might not comes in user life style .
We Proposed a Customer Clustering Algorithm; customer will be segmented based on the transaction that they perform. Based on the Customer transaction detail will generate the Customer purchased tree and Product tree the leaf node product in the purchased tree would the product that the user has purchased, and compare that tree with the product tree to identify the user point of interest in the E-commerce application.
When the User Come in application then based on the purchased tree and product tree the product will be recommended to the user. The Recommended Product are the most popular product in the E-Commerce application.
In Order to Promote Code Item (The Product with less Popularity) among the user, Customer is Categorize into two categories as Normal user and Innovator (Who found that type of Cold Product). The Normal user would became innovator based on the behavior in the E-Commerce Application their Activeness and the number of product they view and time spend for any leaf node based on these category the innovator are found . Once Innovator found any cold product in the Application and found that item to will be useful then that product will be promoted to the Group of Customer whose purchased tree are close the to product tree that the innovator found.