Improving Performance of Search Engines Based on Fuzzy Classification

At first glance, the service search-engine seems very useful and faultless, but by the more careful examination one may notice weaknesses in this search results. One of these weakness is that the result pages, which the search-engines offer is sometimes without content and sometimes have no relevance to the field that user had in mind . On the other hand many of quality-pages have no place in the search results. This paper advises search-engines to hand the job of decision making about the content of web sites to users, because humans are very much faster and have a very lower rate of error and can decide about the usefulness of a website with more justice. In the proposed algorithm which is based on fuzzy logic, we try to use parameters such as speed of mouse movements, scrolling speed, standard deviation of horizontal position of mouse and the time spent by user in each page to evaluate the extent of user's satisfaction with the page content. This ppaer describes the surveys conducted and then analyzes of the fuzzy variables, fuzzy sets and membership functions. Finally, discusses the benefits of the proposed algorithm.