The current explosion of information has made use of search engines inevitable. Search engines are also trying to provide better responses to users. In this paper, an extended version of PageRank algorithm will be presented. In the proposed version, first the pre-processing operation is performed on the log file that includes data cleaning, separation of users and identification of sessions. To delete inappropriate inputs from the log file, data clearing will be used. Time-based methods were used to identify sessions and consider the pages as a session. Interests of users in pages are derived based on two criteria: page visiting frequency and duration of stay on page. Then, inspired by the bee colony algorithm, accuracy was improved. The results indicate that the proposed method has less errors compared to rival methods.
展开▼
