Web user log mining for Web retrieval

As information on the Internet expands rapidly, we can get more information than before. However, how to find user-intended information from the Internet including text, images, and video is not easy. In this paper, we use relevance feedback and build user space by an improved Bayesian algorithm to mine the log of user´s feedback to improve retrieval performance. Data mining is used to remove clutter and irrelevant text information, and help to eliminate mismatch between the page author´s expression and the user´s understanding and expectation.