Personal Information Search Method Research Based On Uncertain Information

With the rapid development of modern information technology, the crisis of "informationexplosion" increasingly brings challenges to personal information management. People needto deal with the great amount of information every day because of the explosive growth ofinformation. In the face of vast information world, users are often unable to find theinformation they need quickly and accurately. When users want to find personal files in theircomputers, they possibly meet with difficulties if they can’t remember the exact file name orlocation. And the existing keyword-based desktop search tools cannot help users solve theproblem effectively. By analyzing the behavior of human memory, this paper puts forward thesolutions according to different demands of information search. This article’s maincontributions are as follows:(1)Put forward a log-based framework on personal information search and a method ofautomatically collecting user access log. User access behavior of the individual informationsearch in the prototype system will be automatically monitored by this method. At the sametime, after a period of time of the prototype system operating on some people of the laboratory,a data set for subsequent experiments is got.(2)Put forward the personal information search strategy based on synonyms. By theanalysis of the collected users’ personal access logs, there is an observation that somesynonym words are utilized to name file by users, and the personal information searchstrategy based on synonym relationship is proposed. It specifically includes: this paper putsforward the personal synonym concept model, defines the basic concepts of personal file andpersonal dictionary, the word preference, synonym graph, file keyword vector. Personalsynonym dictionary and synonym graph incremental method, as well as the personal filesearch strategy based on synonyms are proposed on the strength of the basic concepts.(3)Put forward the personal information search strategy based on the referencerelationship. In the process of further analysis of the logs, there is another observation that thereference relationship between two personal files are bound up with their access time,therefore, a search strategy based on reference relationship is proposed. It specificallyincludes: this paper puts forward a reference model based on access time relationship, putsforward and expounds the basic concepts including reference relationship, desktop accessbehavior, order behavior list, order containing, and further proposes a method to identify thereference relationship based on overlapping time and a personal information search strategyupon reference relationship.(4)Implement a system based on the proposed methods. Based on the access logcollection method, identification method for file’s reference relationship and search strategyof synonyms proposed in this paper, a system of personal information management wasimplemented. And above functions are incorporated into this system which is developed by the laboratory members. Through actual use, the effectiveness of the proposed methods in thispaper is verified.