The Knowledge Discovery Of Customer Group's Common Characteristics Based On Cloud Model And Rough Set

Through the analysis of customer data, we can get the common behavior characteristics of customer base. Enterprises can take corresponding measures to provide customers with different and active service according to these characteristics. This method changes the traditional "passive" service mode, and creates more customer value. It can help enterprises to win customers, improve business performance, and increase the competitiveness in the market. The in-depth command of the customer information, changing this information into knowledge, providing customers with targeted active service, improving customer satisfaction, maintaining the existing customers, and discovering potential customers, have earned the telecom industry, the insurance industry and other industries as well as many scholars to pay attention to..This paper's theme is about knowledge discovery of customer base common character. It uses IT technology and management theory and methods, such as artificial intelligence, data mining, knowledge discovery and customer relationship management, to respectively discusses the customer group division theory, customer group similarity measure method, and mining of customers common characteristics. We improve some weak and inadequate points of the mining process, allowing the customers knowledge discovery more effective and reasonable. The main contents are as follows:Firstly, read a large number of domestic and international literatures, and analyse four research areas including the data pretreatment, the theory and algorithm of rough set, knowledge discovery based on rough set and customer relationship management based on data mining.Secondly, discuss the customer groups similarity measure algorithm. This paper introduces the traditional similarity measure algorithm, and put forward to a new similarity measurement algorithm which applies to multiple attribute customers based on the existing algorithms.Once again, the four key links of customer base character knowledge discovery, namely the discretization of continuous numerical, condition attribute selection, the determination of the association rules and the expression of decision rules, are discussed, and put forward the corresponding settlement tactics. Using cloud model theory to discrete the continuous data, using Spearman grade correlation analysis and genetic algorithm to reduce the attribute, using confidence index and support index to filter the decision rules, using decision tree method, to intuitively express these decision rules.Finally, simulate and verify the entire process of customer base character knowledge discovery by use of instance data.