Research On User-marking Based Social Search Engine

Today, with the rapid development of Internet technology, search engine hasbecome an indispensable part of people’s lives. All kinds of search engines hasbecome a powerful tool for information Searching. However, the search engine‘sPerformance is not always satisfactory. When the user wants to find some information,search engines will return thousands of search results, in which only a few or no Pageof them can satisfy user‘s demand. How to understand the intent of the user‘s search,and how to find pages to meet user‘s requirement and then put the most relevantsearch result pages in the forefront, that have became a very important subject in theresearch of search engine. Social search is coming to the fore in such an environment.The user feedback information is used to filter and organize search results in order toemploy the wisdom of the people and participation. The job is not down by simplyrelying on machine algorithms.The definition of the social search engine has not been conclusive. In thisdissertation, we start from the basic prototype of the social search engines. Thedomestic and foreign research prototypes are analyzed. We analyze user feedback,including the perspective of the most user habits, and combined with the web2.0eraof today’s most popular social labels. A new tag cloud display form of search results isdesigned. Users update the labels of the weights of search results, and this informationis feedback to the search engines.Through the feedback form for the user clicks on the label, we propose theconcept of user feedback scores. The study on specific methods and correspondingrealization for how user feedback to impact the final ranking of search results aremade. A new sorting algorithm for search results based on user feedback waspresented. It is combined the algorithm of the Lucene retrieval system. With userfeedback, the second sort can make the final sorting results satisfying the real needs ofusers better.Finally, the experiments based on this system are carried out and the relevantresults were obtained. The results show that this method can work better.