Research And Application Of Data Mining In Customer's Repairing Data Analysis

Data mining is an effective way to solve problems of "poor knowledge in rich data", and its essence is to extract information which is implicit, unknown but useful to people from database or data warehouse. The data mining is considered to be an important branch which has extremely broad application prospects in database. The purpose of data mining is to fully analyze and understand data, and to interpret the knowledge which hidden in the data, and to show the knowledge in a particular model. The theory research in data mining has already achieved remarkable results by researchers at home and abroad; and about application, data mining technology has been applied to many commercial areas successfully, and has got a good business results.The business of company has become networked and global with development of network technology. And the network has become an essential tool in business transaction. Therefore, more and more attention has been paid on the performance of network. So the data mining and analysis of network repairing database has become a best way to find and locate the breakdown origin, and exclude it in time, which has great significance to business and customers. To the Network service providers, data mining and analysis of repairing database help them find the key factors that caused the failure, and remove the failure origin, and also helpful in improving the network performance. The improving of the network performance successfully promotes to establish a good business image on customers; and to network users, a stable, fast and high-performance network can increase their business efficiency, and also increase economic benefits.Firstly, this paper discusses the status and development trends, system architecture, the basic process, the main task, the main application areas of data mining technology; then introduces the basic concepts of decision tree, and make a research on common decision tree algorithm,selection methods and pruning algorithm. The data used in this paper come from some network services providers in Shanghai, and we use SQL Server 2005 as data storage and management platform, and then apply OLAP and DM as main analysis methods, finally, we use business intelligence application platform to develop programs and show the result.The particular procedures of data mining are as follows:first, defining mining task according to the business requirement and object features; second, preprocessing which need to gather, clean and transform the source data, and eliminate noise of the source; third, connect data source, build data source view, establish mining structure and decision tree model on the business intelligence application platform of SQL Server 2005. The mining and analysis based on decision tree classification algorithm has achieved good results. And the results are compared with the actual situations and observations, and some recommendations on network maintenance and improving methods are given at last.