Abstract

With the growing information in the web, ranking algorithms is very important in searching information. Currently, there are two categories of the ranking algorithm based on contentand connectivity. Ranking algorithms which are based on content have low accuracy and recall and also contain the rank spamming problem. Ranking algorithms which are based on connectivity contain the rich get richer problem too. Therefore, in this paper, ranking algorithm based on distributed learning automata was presented, which use pages content’s information, hyperlinks between pages and web usage data to present better results. In this paper, at first, two algorithms for determining the structure of web documents based on DLA was proposed, and then was used for ranking. The obtained results of the simulation proposed algorithm was evaluated with RankCorrelation and P@n measures.

Key words: Ranking algorithms, PageRank, structures of web documents, distributed learning automata.