Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems

The Exploitation and Application of Customer-Action Analysis System on Date Mining

Author LiangYing
Tutor ZuoHaiSheng;LiZheng
School Guangxi University
Course Computer technology
Keywords Data Mining Data warehouse Consumer Behavior Forecast
CLC TP311.13
Type Master's thesis
Year 2011
Downloads 52
Quotes 0
Download Dissertation

Operators of domestic telecommunications industry after several spin-off and restructuring, as well as the issuance of 3G licenses, industry competition and competition users increasingly fierce, how to enhance service awareness, and the development of sales channels and new modes of publicity are facing new problems. In a competitive environment, customer-centric, if you both have a lot of information, but also has advanced analysis tools, will be able to gain advantage in the fierce competition. Data mining from large amounts of data extraction or mining knowledge for data analysis, which found that the potential of information technology. From comprehensive point of view is more in-depth insight into customers, understand customer value orientation, based on this insight through the appropriate channels to provide tailored product packages to the right customers at the right time can help companies to segment customers. This paper first introduces the theory and development status of the data mining needs analysis of the design goals and functional modules, followed by customer behavior analysis system, given the system flow chart to determine the selection of clustering, decision trees, association rules three algorithms for data mining, the principle of the three algorithms in clustering, decision trees, association rules and details. In this paper, the actual situation of the operators in the communications industry, using the K-means clustering algorithm C5.0 decision tree and Apriori association rules applied to customer behavior analysis, data mining, design and customer behavior analysis system. Data mining module design, first used two clustering algorithms are compared, that the K-means algorithm can be a good solution to the given numeric attribute data object clustering problem, often with a local optimum end , the algorithm is relatively scalable and highly efficient, generally on the sensitivity of the input data sequence, the algorithm results more easily understood, also faster modeling speed coincide with the characteristics of the communication use existing database obtained K-means The algorithm is more suitable for customer behavior analysis conclusion. The K-means algorithm is used to subdivide a communications operator customers for example, described the implementation process of the algorithm, the results were analyzed, and the algorithm has been improved, reducing the K-means algorithm because of its algorithm which may occur in the initial cluster centers randomly selected local minimum at the convergence of the possibility to improve the clustering effect of the algorithm. Second, the system also uses the Apriori association rules for data mining, whether long distance roaming package products for bundling, for example, elaborated on the implementation process of the Apriori association rules, the results were analyzed, and the algorithm has been improved, due to the presence of Apriori association rules repeatedly scan the database to check the candidate set to lead to low efficiency and pattern matching, improved Apriori association rules only needs to scan the database once, greatly enhance the efficiency of the algorithm. Third, the article will relations and the development process of the decision tree algorithm ID3, C4.5, C5.0, CART introduced C5.0 decision tree algorithm is more suitable for the customer spending behavior analysis, the system will be C5. 0 decision tree algorithm is used for the screening of marketing case target customers, customers to long distance package screening, for example, elaborate on the implementation process of the C5.0 decision tree algorithm, and the results were analyzed and summarized. The practical application of data mining technology can achieve good results for the analysis of customer spending behavior. Help corporate decision makers insight into customer behavior, so as to achieve the purpose of improving corporate profits. Finally, customer behavior analysis system based on data mining mart summarized.

Related Dissertations
More Dissertations