Research And Application On Intelligent Data Adapter

With the rapid development of internet informationize, the different departments in enterprises, schools, government and other social organizations all have build their own information management system to satisfy their domain business needs. These systems all have different implementation model and system architecture, and they work independently for a long time. They formed the information island gradually, and produced a large number of heterogeneous data. But as the widely usage of internet cloud computing technology, the platform change of information resource has become a development trend. As result, the enterprises, schools, government and other social organizations have large amount of heterogeneous data to processing, building intelligent data adapter is an effective method to integrate heterogeneous data. Intelligent data adapter can not only apply on different business system, but also can be very efficient for data transform, loading, storage and abnormal data processing.the main work and achievements of this paper are as follows:(1) In order to make generality in intelligent data adapter, this paper studies the expert intelligent data adapter model, extraction and store expert knowledge for business system data transformation, help intelligent data adapter for dynamic data dealing. Based on the study and comparison of deductive reasoning, inductive reasoning, default reasoning in expert system, selected and improved a better deductive reasoning method, which could be applied in expert intelligent data adapter model for dynamic data transformation and processing.(2) In the information platform change process, once the data turned into abnormal or got lost, this will impact the platform service. This requires that the intelligent data adapter can not only transform data, but also need to dealing with abnormal data, to ensure the reliability of the platform. This paper forward the method of using bayesian network and union tree reasoning in intelligent data adapter, for probability reasoning on abnormal or lost data and to fill the data. According to probability of bayesian network joint tree reasoning, we pick up parameter learning to construct bayesian network.(3) In the study and design on intelligent data adapter for college electronic monitoring platform, we summarizes the basic system flow of intelligent data adapter, the architecture with business system and information resource platform, the representation and extraction of expert knowledge, and the specific process to construct bayesian network from prior data set using parameter learning.(4) Summarize and conclude the work done in this paper, points out some deletions, and make prospect to the application of intelligent data adapter.