Recommendation system is a classical approach of e-commerce application for suggesting the products to the users. But the use of recommendation system is performed in various other applications such as caching and pre-fetching, lead generation, CRM systems and others. In this presented work an efficient recommendation system is proposed for design and implementation. The proposed recommendation system inherits the properties of web usage mining and web content mining for recommending the most relevant user next web page for user. The web usage mining evaluates the user behavior and the content mining extract the user interest. In first step the web navigation history is analyzed for obtaining user web browsing behavior. Therefore a list of web URLs are extracted from the web access log. In next the user current search requirement is considered and correlated with their past navigational pattern. In final step the user query semantics are measured and the rank based recommendation is produced. The implementation of the proposed technique is performed using JAVA technology. In addition of that for demonstrating the superiority a collaborative filter based recommendation system is compared with the proposed approach. According to the experimental evaluation the proposed technique is found efficient and accurate as compared to the classical recommendation model in terms of accuracy and resource usage.

Recommendation System Web Usages Mining Web Content Mining KNN Correlation Coefficient Data Mining