Study On The Case-Based Reasoning Retrieval Strategy Of Data Mining Algorithm

Now most general data mining systems have powerful functions and provide a large number of data mining algorithms. But these data mining systems are designed for the data mining experts. At the same time system usage is comparatively complicated. When the data mining experts use these data mining systems to operate data mining, the data mining experts personally choose the corresponding data mining algorithm according to actual application. This makes these data mining systems are difficultly used by those users who are not to known well that the data mining technical term and the data analysis technology. The main research target of this thesis is that it makes more people conveniently use data mining system and ensures that data mining system has higher operation efficiency.The Case-Based Reasoning retrieval strategy system of data mining algorithm is established and the case base is built. Every case includes data mining algorithm and correlative information of some special domanial application. In order to ensure data mining system having higher operation efficiency, system does not build the case index tree according to traditional method, but describes the framework of case organization based on fuzzy quotient space model. Integrating with statistical heuristic search, correlative case is retrieved in the case base. Because of avoiding carrying out comparison to all case in case base, it reduces case searching range and retrieval amount of work and it improves retrieval efficiency and accurate rate at the same time.When user brings forward the relative service request, service request is analysed and correlative case is retrieved in the case base. Then system compares data mining algorithms which extract from the correlative case and choose the best data mining algorithms to carry through data mining. Finally system exports data mining result by picture and educes rationality explanation. Taking the bank customer missing prediction analyse as example, it validates veracity of the Case-Based Reasoning retrieval strategy system of data mining algorithm.On the foundation of researching general data mining systems, the Case-Based Reasoning retrieval strategy system of date mining algorithm which is applied artificial intelligence relevant method and thought is established. We hope that our work can provide certain theoretic basis for researching and exploring general data mining systems.