Application Research On The Technology Of Distributed Data Mining On Marketing

Face continuously worse competition of society ,each enterprise continuously probe how to sell more products.With the data of the enterprise doubly adding,and the way of data processing changing from transaction processing into analytical processing which can support enterprise decision,the database management system that play a great role in the enterprise in the past cann't meet the modem enterprise of basic demand to data processing .Therefore,we must find ways to automatically analyze the data ,to automatically classify it,to automatically summarize it,to automatically discover and charaterize trends in it,and to automatically flag anomalies.So we begin to study and apply the data warehouse technique, multidimension analyse technique and data mining technique.Firstly, this thesis make deeply analyse and demonstration in data warehouse > multidimension and data mining.In the data warehouse. From the definition of data warehouse complete demonstrations about data warehouse and data organization is made.In the mutidimension. Related concept ahout multidimension analyse is explained. Multidimension and online transaction processing analyse are made a contrast.Ih the data mining. Data mining algorithms are introduced and compared. Secondly,In the thesis a most short circuit method of extracting data from traditional database is introduced, by which we could successly ex tract data and build a data warehouse. Meanwhile, Operations of multidimension about slice,dice,pivot,roll-up,roll-down is deeply analyzed,and marketing data is analyzed and processed in the data warehouse. Thirdly, Deeply studying association rule mining of data mining. Introducing two optimized method of apriori algorithm.The first is aprioriTB algorithm,which make the Apriori algorithm more efficient by reducing the size of scanning database.The second is TID algorithm,which make the Apriori algorithm more efficient by reducing the times of scanning database. Finally, Introducing method of classfication and clustering. Studying the classfication of customer's type through decision tree, and studying a new clustering algorithm by the combination of grid algorithm and density algorithm.At present,there has a big gap between the native and foreign on this theme.ln this aspect the native studying was later.So far,in the native there only are several enterprise has built their own data warehouse.The step that finding out useful decision information from a large amount of data has been not realizedThese application has a good outlook in the use of our country's retailing.It not only could help enterprise to more efficiently manage customs ,but also could stimulate consumption and help enterprise more easily changing the latent customer into loyal customer.The thesis realize the data extraction from traditional database by the most short circuit method of operational research.In the aspect of association rule mining optimizing Apriori algorithm from two aspect,and completely analyzing and realizing the two algorithm. In classfication making decision tree mining about the customer's type .In clustering completely analyzing about clustering algorithm based on grid algorithm and density algorithm. But there are some shortcuts in this thesis.I hope to keep up studying the thesis at work later.