Knowledge Discovery in Web Usage Patterns Using Pageviews and Data Mining Association Rule

Abstract

Web browsing and navigation behavior of web users are assessed with web usage patterns. The browsing details that are stored in weblogs consists of the fields namely IP Address, Session, URL, User agent, Status code and Bytes transferred. The three categories of Web mining are Usage mining, Content mining and. Structure mining which are used to analyze usage patterns, content searching and locates the link structure, navigation pattern of the web within the user respectively. The primary data sources for web analysis are server logs, access logs and application server logs. The two most used data abstraction are page view and session abstraction. This work creates different types of page view abstraction and analyzes the web patterns to find the navigation behavior using data mining technique FP-Growth algorithm that generates itemsets and rules with the user accessed pattern. Frequent sets are generated and then association rules are formed with minimum support, confidence. Variant user minsupport values are applied with the dataset to analyze the number of frequent itemsets and association rules for each support value. The proposed analysis creates a multi view of web analysis with variant pageviews and improves the existing FP-Growth algorithm by producing more frequent sets and best rules. This leads to personalize the content of the site as per user needs. This research work is implemented and the result is visualized in Rapid miner tool.