The Application Of The Data Mining Technology On Quality Management Of Tobacco Industry

The 21st century will be the quality century,the quality competition will become the main melody of the market competition. How to strengthen the quality control,the promotion of product quality level,enhance the economic efficiency of the enterprises,already became the most important question of the enterprises. One of Total Quality Management'basic thought is that everything is based on data. With the development of enterprise information project, the database technology as well as database management system widespread application, the enterprise has accumulated a lot of quality datas. People hoped to carry on deeper analysis with these data, as to assist the enterprise to carry on the management and decision-making. But the traditional statistical analysis method is unable to satisfy people's request. They urgently need new technology to withdraw knowledge from these magnanimous data. The data mining is the method to withdraw the concealment, unknown, but latent useful information from the massive data. The data mining as one kind of new policy-making support technology comes from the artificial intelligence. It is a front topic of information science, also question which gathers many research area theories and the practice. Comparing with other domains' technologies, the data mining provides one kind of effective mechanism to discovery the latent relations. In international,data mining mainly applies to the analysis of customer relations management system and the biology and so on .In our country the application of data mining technology has made encouraging progress,for example the Baogang Steel Corporation use the quality analysis and data mining technology to realize quality diagnosis. It has entered new stage of the application of commercial intelligence technologythe .The research about application of data mining to quality control of the tobacco are very few. In this article I do some research about it and carry on the analysis with the methods of connection rule and clusting, draw some useful conclusions. The practice proved that the connection rule and cluster analysis are very useful to the forecast of the quality problem. It can provide policy-making, enhance the product' quality, reduce the quality cost, promote enterprise's core competitiveability.The significance and current situation of tobacco quality control has been introduced first. Propose data mining method in the analysis of existing method' limited foundation, introduced the association rule and cluster analysis. Propose the FP-growth algorithm in view of the Apriori algorithm' shortcomings of scaning database many times. Then and gathered a kind of analysis using the FP-growth algorithm hascarried on the excavation to the quality feedback, the cigarettesense organ appraisal and chemical composition,and has established the B/S structure qualitycontrol information system with the JSP technology,including datamining sub-module. Finally it carried on the summary to this article work,carried on the forecast to the future work.