Research And Application Of Association Rull Mining Algorithm In The Data Mining

Data mining is to reveal the implicated but useful information from massive, noisy, fuzzy, incomplete dataset. Its essential target is to extract valuable information from the large-scale database. Association rule mining is an important branch of data mining, which mainly use to find relevant contact. Because the form of association rule mining is succinct, easy to explain and understanding, and may catch the data effectively the important relation. Now the question of association rule mining from the large-scale database has become the most active, matures, importantly research content in the data mining.In this thesis, we research and analyse the data mining technology, especially the association rule mining technology. Based on the previous research, we put forward corresponding algorithm of mining association rules for the problems which has found in the research process. The thesis mainly includes the following four aspects:First, the data mining technology, the association rules mining technical are analysed and researched. We introduce the basic concept of data mining, and classifies and deduces and summarizes the process of data mining, the application field of data mining and the common technology of the data mining, the domestic and overseas research situation of the data mining in the paper. Meanwhile, we expatiate the basis concept of the association rules by the numbers, and deduce the classification of the association rules and the basis steps of the association rules. We also research the classic algorithm Apriori of the association rule, the improving method based on of the Apriori algorithm and the FP-growth algorithm of no candidate mining of frequent item sets.Second, we have researched the maximum frequent item sets,and proposed an DMFIA-D algorithm for mining maximum frequent item sets based on FP-tree.We explained the algorithm through the example,and validated the superiority and expansibility of the algorithm.Based on the research of the concept of maximum frequent item sets and the existing maximum frequent item sets algorithms, We propose an algorithm for mining maximum frequent item sets based on FP-tree, which algorithm improves FP-tree structure, and makes full use of FP-tree structural features, and uses bi-directional search strategy. The bi-directional search strategy means that the top-down search the candidates item sets of the maximum frequent and the bottom-up count or cut the candidates first, and finally make sure the maximum frequent item sets. Because of cutting down the candidates, so it reduces the time of the algorithm mining,.and inceases the efficiency.Thirdly, in the paper we also research the incremental updating algorithm FUP, and analyse the FUP algorithm, and propose the advantages and the disadvantages of the FUP algorithm. We provide a new algorithm MFUP which algorithm based on temporary table for mining association rules. The MFUP algorithm made full use of the old data mining rules and reduced the times of scaning the database greatly, thus the data mining efficiency increased. The example and the experiment in the paper shows that MFUP is better than FUP.At last, we study the problem of the application of mining maximal frequent patterns algorithm DMFIA-D in the analysis of the supermarket system. Mei Jun(Computer Application Technology) Supervised by Zheng Gang...