A HYBRID PAGE RANKING ALGORITHM

Abstract

The web has now become a very huge network. With the ever increasing pages and users on the web it has become very enormous. The challenge by this is that the ordering of information has become a very difficult task. The modern user queries the Internet. The user needs the information. It uses a search engine for this purpose. The task of the search engine is to answer the user query in very precise and fast manner. There have been very large number of algorithm for ranking the documents on the web. The most popular of them are PageRank, RatioRank, INDEGREE, Salsa etc. But the modern day web is even challenging these well-known processes. The web has become a very difficult to manage information. Therefore, we propose a algorithm for processing the information. The algorithm uses content based segmentation to find the occurrences of term in the page. The value given by the processes is used as the initial value for each page’s rank in the ranking algorithm. The ranking algorithm is made to distribute the value of page among its hyperlinks. Also, we use the differential normalization technique to decrease the number of iterations done by the algorithm. The model was implemented and the results were found out be very good. The system is particularly useful for peer – to peer networks peer or domain based searching which have a small number of nodes, but with better computing environment it can be easily scaled to the web.