Integrating performance of web search engine with Machine Learning approach

Todays diversified user query over web search engine for information retrieval; semantic information for relevant web document on web has been plethora of web search research. A lot many web search engine developed based on semantic meaning like ontolook, swoogle etc., for finding relevant information, which helps to find user based semantic meaning related documents. The concept of semantic similarity or semantic information widely focused in many important fields such as Machine Learning, Artificial Intelligence, Cognitive Science, Natural Language Processing and Web Information Retrieval etc., Traditional web search engines and semantic web search engines relates user keyword with terms, entities, texts, documents which have semantic correlation with user query. Both search engines does not use images within web pages to find more relevant information. Now in this paper we have formulated a web document integrated ranking method based on text semantic information and image based object matching information. This integrated approach presented in this paper does not depend upon semantic information of user query but also consider image appearing within web pages to find more relevant information. Approach proposed in this paper includes finding semantic information using ontology based meaning of user query and feature based object matching over image to find image matching score. In proposed approach combined use of ontology based semantic information and image based object matching score will improve web document ranking.