Research Of Mining Frequent Pattern Algorithm And Its Application In Large Database

Along with the information technology high-speed development and WWW applications, massive amounts of data have been continuously collected in the database of many application areas, which contain much useful patterns, and it is very important to find the hidden and previously unknown information for these areas. Data mining aims at the task of the above work.Association rule mining is an important branch of data mining that it has obtained many valuable results. The aim is to discover previously unknown, interesting relationships among attributes from large database. Due to its simple form and being easy to understand, association rule mining has become a hot topic in data mining.The step of mining association rule including: transform the database into transaction database. Second use the suitable algorithm to mine all frequent patterns from business database. At last produces the valuable association rule by the frequent pattern. Second stage is key. It will decide the accuracy and efficiency of association rule mining. Most of the research attention is focused on efficient methods of finding frequent itemsets. The existing frequent patterns mining algorithm depend heavily on massive computation. And then cause high dependency on the memory size, when tuning parameters, it will cause repeated I/O, and not sufficient for mining extremely large database. Therefore seek not depend on repeated I/O scans and less reliant on memory size become the content of this paper.Through the research, we find there are two factors effect the efficiency of frequent patterns mining algorithms: 1 the transactions layout in database is unreasonable; 2 the problem of algorithms. Therefore this paper will do improvement mainly from this two. The concrete work manifests in following three aspects:1,Has researched the sensitivity analysis method and its in the data mining application. On this basic, we proposed the method that unifies the sensitive analysis and the neural network model apply into the interest mining and profit mining of data mining. Finally produces the mining step.2,In view of traditional transaction in database layout, horizontal layout and vertical layout deficiency. This paper designs one kind of new transaction in the database layout-overlapping layout. This layout can reduce the number times of I/O repetition scanning, Especially when the parameter change frequently. Thus enhance the efficiency of search frequent pattern.3,This paper proposes one kind of new frequent pattern mining algorithm-QFP algorithm. This algorithm has used the data store structure of crossing layout way. first establishes a tree for each frequent item (the QFP tree), then mining each tree according to the condition, until discovers the frequent pattern that conforms to the condition. This algorithm can reduce the number of condition tree, the dependency of memory space, and the computing time of CPU. Thus enhancement the efficiency of association rule mining.