The Visualization Study And Implementation Of Data Mining Model

Now, We just live in the age of network, communication,computer and internet are changing the people's life and the whole society. Lots of information brings some problems as thought they bring much convenience to us. People begin to think that how to find the useful knowledge in the information sea, and keep from submerging in it? Facing these problems, Data Mining and Knowledge Discovery technologies have appeared.As the rapid developing of the database's technique and the wild application of the database manage system, lots of data are accumulated, there is much more useful knowledge behind them, people want to analysis these data in higher level and use these useful data. The database manage system can input, query, statistic the data with high efficiency, but it can not find the relationship or rules among these data, and can not forecast the tendency of the future data .Data Mining is the result of the extended studying of the database techniques .Data Mining technique is one of the new fields, it combined the database techniques, artificial intelligence and statistic analysis techniques and so on, the main peculiarity are that they draw out, transfer, analysis and other modal dispose, and find the pivotal data to help the commercial decision. Data Mining technique upgrade the database technique to higher level, it can not only query and ransack the past data but also find the potential relation among these data, so quicken the transference of data. The usage of the Data Mining technique is to find the potential information and knowledge from lots of the applied data, these data are often incomplete, fuzzy, random and so on.Decision Support System is developed from the base of the management information system, now, developing the DSS is base on the Data Warehouse techniques, and become a set of the solution scheme with the On-line Analysis Process (OLAP) and Data Mining(DM). Data Warehouse supply the data sources to the OLAP and DM. The knowledge mined from the Data Warehouse are showed by the chart, visualization, analogous language and so on, but not all of them are useful, so they must be evaluate, selected, validated, and put these useful knowledge to the knowledge database. According the type of the knowledge, they are classified to summary, relation, class, clustering model and so on, these knowledge can be acquired by relevant algorithms.The paper is finished in the INFOHOLD software company. During work in the company, author take part in the developing work of the Decision Support System of TongHua steel company (TGDSS1.0). The software system are composed of two software sub-system, one is the OLAP software sub-system, the other is the DM software sub-system. The main function of the OLAP is to explore the commercial data momentarily, create the all kinds of on-line reporters, release and send kinds of data files; the main function of the DM is to analyze the commercial data and supply the decision information, involving the function such as pretreatment the rude data, creating the mining model, presenting the knowledge in visual mode and so on. The whole software system integrate lots of techniques such as DM, OLAP, VISC(Visualization in Science Computing), WEB. The system structure are composed of four layers , including the Data Warehouse sever, Data Mining sever, WEB sever and user interface.The DM software sub-system is based on the core Data Mining technique of Microsoft, it make the SQL Server Analysis Services as the Data Mining sever, and build the visual mining model using two type data sources, one is the relational database, the other is the multidimensional data cube; then send the visual result to the user. Dealing with the DM software sub-system, the paper will expatiate the relative content of it. The author's main work as follows:1) first, illustrate some conceptions of DM, the structure of mining model, mining algorithms and so on, then study the theories of mining algorithms and mining model deeply, according the real operation put the decision tree algor...