Based On The Academic Management Of The Campus Network Data Mining System

Data warehouse and data mining is one of the most active parts in research, developing and application of database. Data warehouse is to collect an organization's historical data to a central database so that it can be conveniently processed. It is used to support decision-making and is subject-oriented, integrated, time-variant, nonvolatile. It transfers data to information to support structured ad hoc query, analysis and management decision-making process. Data mining means a process of nontrivial extraction of implicit, previously unknown and potentially useful information from data in database or data warehouse. Data mining provides good technology support for data warehouse-based decision support system, and data mining tools can directly mining in data warehouse for discovering potential knowledge.The research object of this paper is to construct a Ian-based data warehouse for education management, and develop data mining tools based on this data warehouse and the data used for mining is student's examination scores. Data mining tools contain three main modules: data preparation, data mining, result presentation. These three parts reiterate in practice to accomplish data valuation and get valuable knowledge for analysis and decision supporting.In data mining module, this paper introduce four data mining technique: association rules, Sequential Patterns, classifiers, clustering, and mainly research association rules and a new technique through combining association rules and clustering. In association rules research we discuss classical thought and method of d association rules and put forward our improvement: first to add interest threshold to keep the accuracy of the rules, second to use special data structure to improve the speed of the data mining. In combining technique research we analyze the drawback of using association rules to mining the whole data, put forward first use clustering to make the whole data become several classes orclusters, then use association rules with particular perspectives, then research how to rapidly update the association rules after clustering, and put forward new data mining technique.In result presentation module, we discuss natural language, graph and grid presentation involving the association rules. To network users, we develop a tri-layer system of browser/server and middle ware, which make network users can through network to browse the data mining results.