Visualdm: Olam-based Visual Data Mining System

The Visual Data Mining is made up of the technique of data mining and visualization, which is derived from Visualization in Scientific Computing. OLAM(On-Line Analytical Mining) integrates OLAP with data mining, which mines knowledge in multi-dimensional databases or data cube. OLAM makes full use of the advantage of OLAP and data mining, and overcomes the shortages of them. In this paper, the technique of visual data mining based-on OLAM can integrate the tools of data mining with the tools of visualization, and allow user join in the data mining process. All in all, OLAM can improve the quality of data mining and help user understand the result of data mining.Visualization of data mining can make users evaluate and supervise the process and result of data mining, can import users' fondness in data mining early times to reduce the useless results.Because there aren't successful data mining cases in the Police system, this paper propose a new system architecture, integrate data mining with Police data, and establish models according to the Police vocation specialty, mine the rules about cases and suspects from the database to realize the real data mining.Based on "PDSS", aiming at cases and crimes data, this paper add data mining and visualization and explore data mining algorithms(including association rules,cluster,decision trees).Based on OLAM,this paper design a VisualDM system including data mining, visualization and OLAM. it combine OLAP and data mining.user can participate in data mining process so as to change its primary phase.This paper design and accomplish a system of data mining Based-On OLAM which includes four aspects: operation requirement, system architecture, function design, module realization; including three modules: data sources, data mining, and data visualization . According to the study background, the visual components include Association Rules visualization, clustering visualization, decision tree visualization.