Application For Data Mining Based On Decision-tree Algorithm

With the fierce competition between companies, not only the simple way to handle and manage data from database can satisfy people's needs, but also to help manager make decision through integrating many channels of data. On the other hand, with the development of computer technology, the volume of database is up to TB level. Large-scale of data not only make user hard to operate, but also behind these data are valuable information which can be used to make decision. Data mining technology is to be developed and progressed under this background. Data Warehouse and Data Mining are new technologies for decision support, which are increasingly developed within these 10 years. Data Warehouse is to integrate comprehensive data in enterprise for Data Mining, which helps Data Mining focus on core processing. At last, they are correlate-developed.Based on the project of Large Database Pre-research, after learning data warehouse and data mining technologies, this paper designs data mining model under SPSS Clementine environment using C5.0 decision-tree algorithm. Because of the specialty in this project, this paper tests the correctness of data mining model using Foodmart2000 database as test data source. This paper is mainly on the design of data mining model and configuration of parameters of all nodes. After data mining, the result is analyzed to find features of members and give out some useful suggestions for development, reaching the goal of making decision based on information from database. And the correctness of results is tested using model of neural network node.This paper designs and analyzes data mining model for specific theme, implementing theories and algorithms of data mining. Using Foodmart2000 database as test data source, this paper tests the feasibility and correctness of data mining model, providing theoretical and further design basement for application in the field of telecom.