The Visualization Method Of Interests’ Temporal Path In User Personalized Profile

With the rapid increase of Internet information,digital data grows explosively. Leveraging computer technology to find valuable information and knowledge from large amounts of data is an effective way to solve this problem. In order to meet the needs of different users, personalized service can provide targeted content and service for users of different characteristics and different interests.The key of personalized service is looking for the approach that expresses patterns of user profile accurately and effectively.The interests of users change with time and it will results in the problem of drift of user interests.This problem will lead to a decrease of effectiveness and reliability of personalized services.So using the time information to complete the expression of user profile is a viable and necessary option. In response to this problem, this paper discusses the analysis and processing of time information in the process of user profile establishing.Combining the research paths of scholars as well as representation and discovery trends of research topics of scholars, we do some relative application researches. And the major work includes:1)According to the time information of user profile,we propose the methods of weighted keyword time gram and weighted co-occurred keywords time gram which combines the keywords’semantic information and time information.And it provides the data structure foundation for user personalized profile.2)Based on the first step,this paper focuses on the research of the calculation method of time gram. We introduce Ebbinghaus curve and Ebbinghaus function for calculating the weight of the keywords because of the characteristics that scholars focus is changing over time and introduce the impact of the Ebbinghaus function for the time information of user interest profile.This paper also discusses the calculation method of the co-occurrence keywords weight through combining directional affinity formula.3)We propose a method of keywords temporal path.This method divides keywords by time period and then gets keyword temporal path according to filtered iterations algorithm of co-occurred weight of co-occurred time gram of weighted keyword in different time period. And we get the keywords temporal path finally. The process of keywords presents the change of user interests, the keywords time sequence path can be called user interests’ path.At last, according to the basis of data of Wanfang database, this paper does some relative analyses for relative experiments. Through visualized display techniques for time sequence path of key words, the experiences display a new and effective identification method for discovering scholar research path.