Ranking Web Pages With An Approach Of Mixing DistanceRank And HITS

With the rapid development of computer applications, Web scale showing explosive growth. Search engine as indispensable tool for people to surf the web, has been an unprecedented development. In order to better improve the search engine functionality and performance for people to provide more convenient access, as well as to better achieve their commercial value, more and more people began to put into a search engine. The searching and ranking algorithm which is the core algorithm of search engine is becoming hot. To be convenient to address, the searching and ranking algorithm can be called ranking algorithm for short.In this article we will discuss ranking algorithms of search engine. Although there are many ranking algorithms, they have many advantages and disadvantages. Taking it into account that the advantages and disadvantages are different for various algorithms, we plan to join two of them together and make it a new algorithm. The new algorithm will have two kinds of the integrated nature of the different methods and can be as far as possible avoiding weaknesses. HITS algorithm, as a query-dependent ranking algorithm, is becoming more and more popular, and DistanceRank algorithm based on reinforcement learning as a global off-line ranking algorithm has been introduced in the near past, which has excellent performance and very good potential of development. Based on the consideration above we improved DistanceRank to be a query-dependent ranking algorithm which is called QDistanceRank, and we mix this ranking algorithm with HITS and obtained a new algorithm, we call it QDRHITS (QDistanceRank and HITS Algorithm).This algorithm is implemented in Java language, and we turn to the classical PageRank algorithm to evaluate the performance of the two algorithms. Experimental results show that, the ranking quality of the two algorithms are better than the classical HITS algorithm.