Design And Implementation Of Data Mining Algorithm Package Prototype System Based On DMX

With the rapid development of computer information technology in recent years, data mining technology plays more and more important role in activities such as enterprise management decision. Mass data within many major enterprises or in their interaction with business partners need be transformed into useful knowledge and information by data mining technology and platform. Data mining technology sprouted in 1980s, and after rapid development in 1990s, kinds of data mining platform technologies, business software data mining algorithms emerge in endlessly, especially the emergent and constant perfection of data mining business software offers major enterprises more advantageous criterion when dealing with enterprise management and decision activities. When discovering inner patterns, disciplines and tendencies after continuous mining on mass business data, enterprise decisions in market becomes more specific, and at the same time unnecessary strategy risk reduces greatly.However, while major enterprises existed enjoy the fruit of newest research and application of computer information and data mining technology, minor enterprises cannot get related system and software technology which match their business requirement. Data mining softwares such as Warehouse Miner and Enterprise Miner, which are leaders in their class and highly integrated, is too expensive and inappropriate to them, and minor enterprises require much on software architecture flexibility, openness and platform independence, with the development and transition of enterprise strategy. Therefore, how to design and implement a data mining software that has high openness, platform independence and expansibility features to satisfy minor enterprises'requirement about inner large scale data processing and knowledge discovering in open source platform is also the main topic of this paper.In recent years, companies like Teradata, SAS and Oracle come up with and continuously perfect their data mining software, and these are almost perfect in function and performance, and they satisfy requirement of users very well from data warehouse construction and the application of front end data mining tools. However to unify and standardize a data mining language becomes a research aporia and hotspot of industry. Among them, specification OLE DB for Data Mining raised by Microsoft and DMX data mining language have been taken as representatives, which reflects the maturity of the standardization of data mining technology.This paper starts with related rationale, elicits OLE DB for Data mining specification from data mining principle, method and technology of data warehouse and the development history of DMX language and states present hot research subject on data mining technology. Then based on the actual requirement of data mining softwares by minor enterprises, this paper accounts the features and advantages of the technological measure and framework used in designing and implementing a DMX language based data mining algorithm package prototype system. After that it expounds the process of designing and implementing the DMS language based data mining algorithm package prototype system from the angle of software design approach. At last, this paper analyzes the advantages and disadvantages in main structural property by contrast among this prototype system and the data mining platform of Microsoft SSAS, and prospects the goal of the further work and research content, in order to serve as a catalyst for work on data mining platform and technology in all respects.