The Study Of Data Mining Based On The Statistics View

From the end of 1980s, when data mining was known for the firsttime, more and more experts from distinct areas are interested in studyingdata mining. This thesis studies data mining based on the statistics view.The thesis includes seven chapters apart from introduction.Chapter one: the theory system of data mining from a statisticalperspective. By comparing data mining and statistics in many aspects, thethesis puts forward the theory system of data mining from a statisticalperspective. So we can understand data mining and statistics moredistinctly.Chapter two: summarizing statistics methods of data mining. First,the thesis discusses three issues about data mining: data, attribute typesand functions. Then it summarizes statistics methods about associationrule,clustering,classifying and regression, and also it improves somemethods from data mining application perspectives.Chapter three: studying statistics methods of data mining deeply.The thesis not only studies methods of character mining but also distancefunctions and resemble coefficient of clustering. Chapter four: studying qualities of data mining. The thesis thinksthat qualities of data mining include three parts from whole process: thequality of data, the qulity of data integration, the qulity of data analysis,then it studies sone methods to improve qualities of data mining from thestatistics view.Chapter five: realizing the data mining antetype system. The thesisstudies design elements of the data mining antetype from applicationscenes,users,process models and models expressing. Then it designs thedata mining antetype system LavaMine. LavaMine has threecharacteristics: flexible, expansibility and encapsulation.Chapter six: an example of data mining. The thesis performs datamining on a database as an example. The database is about ZHEJIANGprovince Unicorn colorful ring customers.Chapter seven: summing-up and study expectation.