The Implementation Of Portable Personality Service Technology Based On Ontology Of User Profile

With the continuous development of information technology, web applications have become an indispensable part of people’s lives, while the explosive growth of information leads to information maze. Personalized service utilizes the user preferences and behavioral characteristics of interest to build user profile, in this way, different information and services will be recommended to different users, then, we can meet the needs of individual users.With the technology of the personalized service becoming more and more mature, more web sites adopt this technology. However, user profiles are often independently dispersed in many different platforms. It has turned out that user profiles can only be applied in their own platforms. In addition, applications are required to maintain user profiles in different platforms, which tremendously decrease users’satisfaction.In order to address the issues above, we design a portable personality service framework based on ontology method, which can share and exchange user profiles between different platforms. This research mainly includes the following contents:1. Automatic construction of domain ontology. This paper introduces the current ontology construction tools and methods. Considering the current ontology construction process still needs a large number of manual operations, we design an improved method for automatic construction of domain ontology based on lexical collocation. The experiment also proves this method is effective.2. We present a robust universal portable personalized service framework, which includes how to obtain user profile, how to submit user profile, how to manage user profile and how to manage the application.3. In order to verify the effectiveness of personality portability, we design a recommendation prototype system in mobile devices. This system can extract current user profiles from bloggers in the science web. Then it can submit user profiles to the portable personality server. In this way, user profiles will be separated and academic literature information will be recommended to the current blogger according to his profile. Experiments show that the satisfied effect can be achieved and also validate that we can share user profiles in different platforms based on the method we proposed in this paper.