Improve Search by Optimization or Personalization, A Case Study in Sogou Log

With the development of information technology, especially when the Web is developing and widespread, the information resource grows explosively. The information retrieval(IR) system becomes more and more important as an indispensable way of accessing the information. Generally speaking the search engine returns results for user community rather than each person so there is also considerable room for improvement. But, the user of search engine cares about the personal results for him. And many personalization strategies have been investigated. We present a large-scale research of Sogou Logs and do the following work. 1) Analyze the performance of the search engine, 2) According to current information on the user logs, calculate the optimal performance of the search engine without personalization. 3) Process the current user logs and calculate the optimal performance of the search engine with personalization. 4) Adopt two possible personalized approaches and calculate their performance respectively. 5) Reveal the relationship between the performance of personalization and the results returned by the search engine. All of our experiments are carried out on Sogou Logs.