基于网格的数据挖掘框架和元数据模型研究-计算机软件与理论专业论文.docxVIP

AB AB STRACT Due to the increased computerization of many industrial，scientific，and public sectors，the amount of available digital electronic data is growing at all unprecedented rate．n坞effective and efficient management and llse of these data,and in particular their transformation into information and knowledge，is considered a key requirement for Success in such knowledge-driven sectors．Data mining is the de-facto technology addressing this information need．But the traditional Data mining systems which run on single machine or local cluster Call not process the distributed data effectively, efficiently．So here give a solution base on Grid computing． The main works of this paper include two parts．Firstly,Survey of the theory and technologies grid computing．After comprising the architectures of grid nowadays，the deficiencies of these鲥d architectures are presented．And indicates that to solve the problems of distributed data mining realize the Data mining酣d is the right way．By analyzing the requirements of the data mining鲥d and comparing the architectures of the鲥ds，this paper proposed architecture of the data mining鲥d．Based on the Globus Toolkit and other open technology and standards,the architecture provides tools and services facilitating the班d—enabling of data mining applications without any intervention On the application side．Critical features of the architecture include flexibility,extensibility,scalability,efficiency,conceptual simplicity and ease of use． Secondly,this paper investigate meta-data schema used for Data mining鲥d．After the actual data mining program(i．e．a batch-style executable)is uploaded on a鲥d server and an XML document(i．e．an instance of the meta-data schema)that describes the program is prepared and registered with the underlying grid information services．Users Can discovery and execution of the program in the鲥d environment easily． KEYWORDS：Grid computing；Data mining；Grid architecture；Meta-data schema CLASSNO：TP301．6 学位论文版权使用授权书本学位论文作者完全了解