Research And Design Of The Off-network User Analyzer Of Mobile Telecommunications On The Basis Of Decision Tree

The telecommunication industry is the pillar industry and the economic lifeline of the country. In recent years, under the support of the nation, telecommunication industries have obtained considerable development. The developmental speed of the mobile communication trade is at a more tremendous pace among them. Nowadays, in mobile communication trade, on the one hand, new users who enter the network increase constantly; on the other hand, the off-network users are also in the constant increase. Compared with the advance of networks and business, analysis of off-network users seems drop to relatively behind, and pre-warning analysis of off-network user is hardly a subject for mobile runners. This analysis has great significance for mobile company, it can improve their ability to make efficient strategies, so as to retrieve those off-network users and to reduce the running cost.On the basis of the current situation of the mobile communication company, this thesis aims to design an off-network user analyzer while taking accounts of the basis fact that this industry has to deal with a great quantity of data, and is characterized with complicated attribute. This analyzer has already been used by the Mobile Communication Company of Heilongjiang Province since 2004 . The main points of this thesis are as follows:On the basis of the data characteristics of this trade, this thesis points out that among the data mining discipline the most suitable model for off-network user analysis is classification, and the most suitable algorithms in this model are decision tree creating and pruning.This thesis, accordingly, proposes and designs an off-network user analyzer on the basis of decision tree for mobile communication industry. Concerning the core part of this analyzer-decision tree creating algorithm in decision tree module, this thesis proposes a new realizing strategy to make the creating more efficient.Furthermore, this thesis puts forward practical tactics for these important components of the analyzer, such as data file store, evaluation of decision tree, and communication of the inside modules.Replying to the requirement of the mobile communication company, this thesis employs this analyzer in the process of the company's off-network user analysis, and lists the practical results of this application. At last, this thesis is summarized by making expectations to various kinds of other analyses that mobile communication industry will make in use of the data mining.