Analyzing User Behavior on E-Commerce Sites by Using Parallel Incremental Forward and Backward Frequent Path Traversal Approaches

Abstract

In recent years so many e-commerce sites of various organizations are emerged around the world. The main intension behind these sites to provide essential goods. The human behavior on these sites vary continuously depends on their interests on new and attracted items which termed as web mining. As going on users may search for different items until they find desired one. To identify the user behavior we need to traverse the path associated. There exist a maximal forward and backward path traversal approach (MFPT and MBPT) which tells human behavior one-commerce sites. In existing approaches user behavior predicted up to 60-70%. To get improved user behavior analysis, in this paper we proposed a parallel incremental path traversal(PIPT)  approach which uses both forward and backward frequent path traversal. With this we can analyze user behavior in accurate manner by generating a FB traversal tree. Compared to existing works our proposed approach shows better performance by means of accuracy and reduced time for analyzing user behavior