Research Of File Search Technology Based On Contextual Information

With the development of storage technology in the past decade, the storage capacityof personal computers is increasing rapidly. According to research, user didn’t like todelete old files to ensure that important data will never be accidentally removed. Thismakes people often spend a lot of time for document retrieval. Traditional file search tooldoes not solve the problem that the user forgets their search keywords. At the same time,using the personalized features to sort the search results is another concern of the user.In order to help user to manage their personal resources more efficient，this paperpresents a context-based information solution. First, this paper improved contextrecognition algorithm. This algorithm use the information of user’s activity at theapplication layer and the file system layer, to establish context relations between files intime and causality, based on analysis the operation of the paste-copy and focus windowchanged. Then, this paper improves the breadth-first search algorithm in Graph Theory.The maximum search level and cut operation based on the relationship between files isdesigned in it. In sort algorithm, this paper improved the parameters of the machinelearning system and put the context information of user’s activity in it. We choice sevenparameters to machine learning system, based on the comprehensive consideration of thecontext information generated by the user behavior and file context information. At last,this paper completed the file search prototype.Finally, this paper makes a series of test. According to functional testing andperformance analysis, with maintaining good precision rate and recall rate, this file searchsystem can provide context-based information extended search service for user morefriendly and efficient management of personal information.