Expansion Of Dynamic Query Based On Query Log

Nowadays, as the rapid development of the internet, collection of the information on the internet is gradually increasing and the growth rate of the information is rapidly developing. Meanwhile, the internet users are increasing. Facing a large number of the information provided by the internet, it is an important research topic that how to get the information that meet everyone’s needs. Therein, the query expansion technique is a very useful method for information retrieval. In the process of user queries, its historical query logs are highly informative. You can get extended words that matching with intents of users’current query based on these query logs.In this thesis, we propose a new query expansion model based on user’query logs. The model is improved based on a exist algorithm in the thesis:query log based query expansion algorithm, two influencing factors are added in the module of the relevance documentation, which one is the length of users’query time interval and the other is the same query documents’feedback for different users. This model can obtain the relevant documentation to provide the better query expansion words as query expansion source, and then a new query can be formed. The new query expansion can more accurately describe the users’intents. Thus, the model can improve the retrieval performance and user satisfaction.We discuss the new method based on query expansion, and then realize a lot of experiments. According to experiments, parameters of the query expansion are calibrated and the quality of query expansion words is improved. Extensive simulations are carried out and confirm the proposed model can improve the retrieval performance and the user satisfaction.