Classifier cum Recommender System for E-Governance using Collaborative Trie

The boom in the government services online has put a great difficulty before the users in selection of the desired web pages on the e-governance portal. This has increased the requirement of such a recommendation system which intelligently satisfies needs of a huge user base. The intelligent collaborative recommendation system proposed in this paper analyzes web logs using web usage mining techniques. It provides such an interface to the existing users where they set their preferences for personalized and collaborative recommendations. Experience of previous users is a cornerstone for recommending default set of pages to na ve users. The use of efficient data structure trie performs dual functionality. This clusters the similar users together which saves the efforts of applying a separate approach for categorizing users. In addition, it conveniently recommends the desired pages to the user. This system dynamically changes the support value of a pattern. This automates the promotion and demotion of a pattern to a group. The hashing technique efficiently finds pages of user interest, in the trie in 0(1) time complexity.