The Research Of Data Mining In The Telecom Business Customers' Behavior Analysis

So far, the use of the communications industry in general to talk about a variety of systems are mainly based business processing systems, It is mainly used to enhance the operational level, to improve efficiency, standardize business processes. The rapid development of business at the same time. The level of information technology within the enterprise has made relatively large construction progress, so that the efficiency of China Telecom's related work has been greatly improved, making service quality and management level of the important role has been brought into full play. After many years of running the system has accumulated a wealth of data, if such a comprehensive analysis of the data can be fully aware of all aspects of China Telecom's business activities, customer behavior analysis to provide reliable scientific basis.This article first introduces the data mining technology in all sectors in the application of data mining technology research led to the status, trends and the existing problems. Then introduced the data mining techniques are often used in association rule mining algorithm—algorithm, analyzed the association rules algorithm Apriori algorithm and the classical algorithm FP_Growth algorithm, these algorithms in a large database application flaws. Then according to the questions raised by research, a new algorithm based on FP_Growth improved algorithm, and comparative analysis of the algorithm and the performance characteristics of classical FP_Growth algorithms, and finally algorithms for new research into the telecom business customer behavior analysis instance.The main characteristic and innovation of the article is below:(1) Data mining technology to telecom business customer behavior data analysis, in the telecommunications industry is still immature, so the theoretical and technical advantages;(2) Application environment for telecommunications companies and demand, the Fp-gorwth algorithm optimization and realization of its performance were tested and applied to practical systems.