Clustering And Case-Based Reasoning In Data-stream Management

Data-stream mining is a new research aspect of data mining. It has become a useful tool for many fields. Traditional databases store sets of relatively static records with no pre-defined notion of time, unless timestamp attributes are explicitly added. While this model adequately represents commercial catalogues or repositories of personal information, many current applications require support for online analysis of rapidly changing data streams. Limitations of traditional DBMSs in supporting streaming applications have been recognized, prompting research to augment existing technologies and build new systems to manage streaming data. Nowadays the main research of Data-stream is sequence query, and the main research of Data-stream mining is research on the result of the Data-stream system. For example, KDC (knowledge discovery in case base) , clustering, intelligence decision and so on. But no system can mine on the Data-stream straightly. It's also reducing the function of the system.Recently, because of the demand in many fields, researchers pay more attentions to the problems of the Data-stream, especially the problems of data-stream mining. Nowadays the disadvantages of data-stream system and the one-pass of data-stream lead that it is hard to get and mine useful data from huge data-stream. The main content of this paper is to review recent work in data-stream management systems, with an emphasis on application requirements, data models, continuous query languages, and query evaluation. And research of the Clustering and Case-Based Reasoning (CBR) apply to the data-stream Management, proposed an effective data-stream cluster algorithm and a data-stream management system which can be used for querying, analysis, mining data-stream.In this paper, we propose a new model of data-stream management system which based on data-stream system, data-stream cluster algorithm and CBR. Firstly, this model clustering the results of data-stream system, then get case from each cluster, and add to the case base that is the preparation for advanced mining. This article expounds the model of data-stream cluster algorithm and data-stream management system. This initial experimental prototype has shown us a satisfactory result.