Research On Chinese Personal Email Management Based On Topic Detection

With the development of Internet technology, various kinds of network application services welled up. E-mail as a means of swift and economic communication is widely used by the users, which results in the email overload. The email overload problem has become an urgent problem to be solved. Recent email clients can no longer meet the people's needs of dealing with emails, so email management technology has become a hot research topic. This paper used the idea of topic detection and tracking for reference and devised a scheme for topic recognition in email. We mined the context relationship between the emails, clustered the emails which belong to the same topic, through that we hope to improve the sorting and searching function in nowadays email client and help users to search the related emails more quickly and conveniently. The main work in the research includes the following elements:1. Introducing the technology of topic recognition and tracking into email managementOne result of our thesis is that we proved that introducing the technology of topic recognition and tracking into email management is feasible.2. Proposed a email address similarity and a title similarity calculation method respectivelyAccording the analysis about the traits of the email text, we proposed an email address similarity and a title similarity calculation method respectively, and those were testified that the performance of these two methods is better than the performance of the existing methods. We also applied two email content similarity calculation methods and through the comparison experiments we chose the better one. We combined the characteristics of mail's title and address into the topic recognition similarity calculation method, and the results indicated that the improved similarity algorithm is better than the arithmetic that just based on the content similarity.3. Finished an Outlook plug-in model based on topic recognition.Finally, we considered the users' operation habits and convenience we completed an Outlook plug-in model based on topic recognition.