Research On Evolution Mechanisms Of Mobile Personalized Information Service System

While the Web and communication network have become main tools of people for acquiring knowledge, the problem of Information Overload emerged. To solve this problem, personalized Information service system aims at providing users information, which is consistent to their interests. Currently, novel frontier, which is referred as Mobile Computing, has become the trend of Web and brought new challenges for personalized service. The working mechanism and the service model of traditional systems are consisted of two main steps commonly. Firstly, acquire user feedback and analysis users’interest during the interaction. Secondly, customize specific contents by modify the value of existed user model and change parameters of service model. However, the effects of the process can not work so well when used in Mobile Computing. Since mobile service takes several new complex factors into account, this leads to the inefficiency of adaptation for user interest changing. Further more, traditional methods depend on feedback excessively, there are rare research focusing on the rule, which triggers users’interest changes. The lazy learning lags behind the interest changing and then provides reactions too passively.This paper aims to resolve the problem mentioned above, dedicates to research the active evolution mechanism, including the organization improvements of components, the optimization of appearance of the system. Beside, we should ensure the feasibility of the evolution mechanisms. This paper made efforts on the orientation of user interests change, system evolution with constrains, the methodology of software evolution. The main researches and innovations are as follows:(1) Research the detailed orientation of the interest changing and implicit rules, proposed methods for acquired interest changing knowledge, which are caused by user cognition, information seeking behavior, and context changing, propose methods to predict user interest and improve the quality of information service.(2) Research the evolution mechanisms of mobile personalization information system. Proposed the methods of the reorganization of contents on server, the content adaptation in terminals and the constraints of these re-factors. Propose an intelligent method for configuration recommendation on mobile terminals. This method can recomment configurations to user based on the current context.(3) With the interaction between large numbers of users and contents, large scale data will be generated. Beside, in real cases, the large scale data is in the form of data stream, which comes from distributed places. Thus, research the data mining methods over the fast arriving data streams, which include the frequent item-set mining and frequent sub-graph mining. Proposed a frequent sub-graph mining methods over data streams, which can handle concept drift by keeping a concept samples during mining.(4) The evolution of software systems, may encounter many problems in software engineering problem, which caused by redevelopment. Therefore, this paper researched design pattern, software evolution, requirement engineering and etc. This paper aimed to combine the user interest changing and software evolution, map the change of user interest changing to the changing of system components, by which we can further guide the development and evolution of mobile personalized information system. Such activities can improve the flexibility and scalability of the system, and have benefits on mobile personalized information system evolution.