Application Research Of Data Mining In Enterprise Marketing Decisions

DSS (Decision Support System) has got considerable development in China, butimprovement and perfection is needed to the theoretical research and practical application of it.DSS plays a very important role in business activities, especially in decision analysis. Not onlycan save a lot of manpower, improve the efficiency of decision analysis activities, reducing theoperating costs of enterprises, but it also remarkable in accuracy and immediacy of informationand in displaying of information with multi-dimensional and multi-degree. A series of informationconstruction has been made by Zhejiang tobacco, in order to adapt to the development andimprove the competitiveness. Designed the DSS based on data mining, with the actual demandin Zhejiang tobacco. Specific research includes the following five parts:(1) This thesis first elaborated the thesis background, the development status and trends ofDSS. Briefly introduced the Data Mining, Data Warehouse, OLAP, core technologies of DSS,and the relationship of core technologies, discussed the disadvantage of traditional DSS and theadvantage of new DSS.(2) Analyzed and designed the DSS based on data mining, with the actual demand of salesanalysis, forecasting, market launch analysis and retail analysis in Zhejiang tobacco, designedthe system architecture, studied and divided the analysis topic, on this basis, functional modulesbeen designed.(3) Designed Data warehouse of the DSS with conceptual model-logical model-physical model, anddesigned the physical table structure for theme of cigarette sales. Discussed the scheme of ETL, and utilizedthe scheme for ETL operating, which joined SQL programming and Datastage of IBM, showed the ETLworkflow provide support for decision analysis and multi-dimensional analysis.(4) This thesis has made the analysis of decision making method on time series, regressionanalysis and decision tree. Establish cigarette sale prediction model with the seasonal predictionand regression to predict the future sale, and provide support for formulating market strategies.Classification analysis on customer using ID3has been made, to solve the problem of customerclassification, and formulat different strategies customer in different classification. (5) Collected, processed, integrated the data for data mining, the work of data preparationhad been completed. The data include: Purchasing, sales, inventory data and information ofretailers, and the retailers are more than960million. The macroeconomic data and populationdata in the recent twenty years. The data from market and business system.