Real-Time tracking of user's interest fields and search result personalization using the national science and technology standard classification

Abstract

As researches on convergence between diverse fields of science and technology are activated, researchers’ interests are shifting to other areas as well. Therefore, the information service that needs to transmit information quickly and accurately should identify the area of interest of the user and reflect it in the search result. In this paper, we propose a method to improve user's satisfaction by grasping the user's interests in real time and reflecting them in search results. This paper proposes a methodology in which a user's fields of interest are identified by tracking the user's search behavior and interest in science and technology information (research subject information, articles, patents, study reports) in real time using the National Science and Technology Standard Classification Code; this information is reflected in the search results. For classification information to determine the user's fields of interest, the medium codes of the National Science and Technology Standard, pre-granted in the document records for science and technology information, are used. The extracted data regarding the user's fields of interest is applied for re-ranking the search results according to the user's levels of interest, using the boosting method. The users ‘interests were different from day to day, and it was found that the documents that boosted the search results by using the S & T classification code are more suitable for the users’ needs. Especially, when the search is performed with a simple word, the method of presenting the search result using the interest field at the top greatly increases the user satisfaction. It is not meaningful for a physician to first present a document containing the word ‘virus’ in computer science. The results of this study can be applied not only to science and technology but also to various information retrieval services in humanities, social and political fields. In the future, it is necessary to experiment with how precisely to track and detect changes in user interest.