Novel framework on particle agent swarm optimization in semantic mining for cross domain web page recommendation

Abstract

The availability of data and information across the internet grows exponentially with time The web acts as a powerful and useful tool for users to access and retrieve relevant information Search engines play a vital role in bringing relevant information to the users on a query basis Web page recommender systems play a vital role in identifying users information requirements and make their browsing efficient Numerous web page recommender systems are built that chooses as well as suggests online pages that are appropriate with users need of the hour In current years there is raising attention in using data mining techniques in the web content to construct webpage recommender systems The primitive data mining techniques are not found to be efficient owing to their incapability to handle the pages that are created dynamically Therefore traditional mining methods need to be often changed in a recommender model In this work initially a Two Phase methodology is proposed for an optimized web page recommendation It is performed based on machine learning from web logs and to recommend users with web pages that are of importance to users interests by comparing it with users historic browsing patterns The web search can be optimized to provide the user with exact web pages from where he can obtain the information of his search goal Finally the result is optimized by assigning new ranks to the result pages The performances of the search engines are improved and thus enabling the user to attain the relevant information he needs The proposed methodology proved to be efficient in terms of the page order which eventually reduced the search time newline