Research Of Data Mining Algorithms Based On Association Rules

In recent years, many people in information industry attach more importance to the data mining technique, which is attributed to the necessary consequence of the conflicting movement between the rapid-increasing data and the poor information day by day. Studying the data mining technique systematically, deeply, roundly and detailedly is an objective requirement for exchanging information in the global. This dissertation systematically, deeply, roundly and detailedly studies and analyses the data mining technique, especially the one for association rules. The main contents are listed as follows:Analyse and research of the data mining technique. The appearance of the data mining technique is reviewed in brief first. Based on the basic concepts of data mining, this dissertation classifies and summarizes the objects of data mining, the fmdable patterns and the common techniques in detail. In succession, the dissertation summarizes, analyses and studies the current status of the data mining Technique in our native country and overseas widely and roundly and then summarizes and discusses its developmental trends and hot research fields. All of the above become the basis for this dissertation.Analyse and research of the data mining technique for association rules. Based on the basic concepts of the association rules, this dissertation classifies and summarizes its species roundly and summarizes, analyses and studies its typical mining algorithms and these algorithms' basic ideas in detail. In succession, the differences among these algorithms are compared objectively and the consequences are illustrated through an example. All kinds of optimized techniques which are designed to promote the algorithm's efficiency are also studied and discussed in detail here and at the same time their advantages and disadvantages, i.e. their merits and defects are analysed objectively. All of the above rationally establish the necessary premise for HY algorithm's proposition and construction.Design, analyse and research of HY algorithm. Considering the defects of typical algorithm for mining frequent itemsets, this dissertation puts forward HY algorithm which is designed to mine association rules and based on the hashtechnique and the optimized transaction reduction technique. First, Considering the characteristics of mining association rules, an effective hash function is constructed and its constructional principles, realizable methods and efficiencies are analysed, studied, discussed and proved in detail and at the same time several new concepts such as radix-scale degree, combination-existence degree, combination-denseness degree and so on are defined too. Second, based on the advantages of the traditional transaction reduction methods and the defects of the transaction reduction technique in DHP algorithm, an optimized transaction reduction technique is proposed and at the same time its operational principles and realizable steps are analysed and studied in detail. Successively the process of mining association rules using HY algorithm is illustrated through an instance. Finally the algorithmic steps of HY algorithm is described in detail.Experimental results of HY algorithm. Based on the fact of generating the synthetic data using Poisson distribution function and exponential distribution function, the performance of HY algorithm and the comparison among HY algorithm, Apriori algorithm and DHP algorithm is experimented. These experiments include the one that compares the execution time using variant synthetic data and variant minimum supports, and the scale-up one that compares the execution time using variant transaction number and variant item number in synthetic data. Finally the results of the experiments are analysed. All of the experiments reveal good performance of HY algorithm.Realization of the prototype system. In this dissertation a simple prototype system for data mining is realized using Microsoft VC++.net and Microsoft VB.net on Microsoft Windows 2000 Server, Microsoft SQL Server 2000 and Microso...