Study And Design Of Data Analysis Assistant System Base On EJB

Data mining has great application prospect in the real world. It's an important tool in data analysis.As data mining technology used in enterprises, two application problems rise, one is the distribution of application, and the other is the need of mining large dataset. Moreover, the limitation of software and hardware conditions in enterprises should be considered. Thus a feasible solution, which is able to meet the requirement of application and environment status from enterprises, is urgently demanded.This paper introduces a solution to develop a general-purpose data mining system that can adapt to the environment of enterprises. This solution also includes the application of approximately mining technology to meet the requirement of mining large dataset. We design and implement the DAA (Data Analysis Assistant) prototype system, then carried out mining practice with this system.As mining large dataset is necessary in real world, through theoretical research and simulating test, we come to the conclusion that there're three main points needed in performing an approximately mining practice. The three main points include supporting algorithms, mining framework and the consideration of approximately mining technology during mining large dataset.The functions of DAA include browsing, statistics and data mining. The core function is data mining. The discussion on DAA system design includes the implementation of fundamental functions and the optimization of the system's performance.Design work includes system structure design, service design and client interface design. As to system structure design, multi-tried application architecture and EJB technology is considered. Service design is the major work, which includes EJB service design and Web service design. While client interface design includes interface design, task description and model description design.On aspect of optimization, we combined several key solutions including coarse-grained accessing method, intermediate data management technology and asynchronous communication technology in mining procedure to improve system's performance.At the end, a typical application example - DAA used in bank loan scenario is presented. The scenarios of building model by analysts and applying model by customers are illustrated. An instance using approximately mining technology is also presented.