Improving Web Usability With Data Mining

Abstract

We study the usability of a web-based online retail store (similar to Amazon.com) using data mining. We record and later analyses the behavior of the users on the web system with the help of interaction pattern mining. We record user interactions for the purpose and later mine those interactions using sequential pattern mining to find most frequently occurring sequences. These sequences help in identifying the most frequently paths taken in the application and thus the most user friendly areas in the application. Application areas not getting much attention are natural candidates for improvement from a usability perspective. We develop an Online shopping store similar to Amazon (V-AMAZON) which will include a subset of products from Amazon and the idea is to record user interaction and then use mining algorithm to predict user behavior.