The Mining Of The Design Criteria Of Forging-die Based On Numerical Statistic

Data Mining can exrtact hidden, unknown and non-ordinary information or models which have potential applications from the large databases or data warehouse. It integrates the theories and technologies of database, artificial intelligence, machine learning, statistics and other fields, being a valuable new field in database research.There are many methods of discovering the regularities among variables in numberial statistics, which is one of three main branches among data mining, but regression analysis is one of the most effective ways in it.Regression analysis, as one mechanism of data mining, is discussedand and an instance of constituting forging-die design criteria using it is given in this paper.The main work is summarized below.(1)The current study situation of the application of numberial statistics in data mining at home and abroad was surveyed. The features of the combination of numberial statistics with database technology are expounded. The data mining technology based on numerical statistic is sumamarized.(2) The relationship between numberial statistics and data mining is expounded. The study history and current situation of data mining are recounted. The definition of data mining is given. The relationship between data mining and related technologies is discussed. The knowledge type found by data mining, functions of data mining, technology commonly used in data mining, data warehouse in data mining and so on are conferred. The work principles of data mining systems are expounded.(3)The basic concepts of regression analysis are expounded. The definition of linear regression equation is given. The algorithm of estimating the parameters in forging-die flash size design criteria and the method of significant test in linear correlation are expounded in detail. The non-linear regression equations, the model of multivariate linear regression equation and its algorithm of estimating the parameters are discussed. The methods of using numberial statistics in drawing up process algorithms are discussed.(4)The algorithms and mathematics discription of flash computing are analysed. An instance, setting up forging-die flash size design criteria by means of the step-by-step regressive analysis method is given. Such problems as the development of the step-by-step regressive analysis software, the constitution of forging-die flash size design criteria, the constitution of forging-die flash metal consumption design criteria and so on, are studied. Finally, the two criteria above are established with the step-by-step regression analysis software used. Then the results were analyzed. The efficiency of the algorithm was verified by experiments.