An Improved Algorithm For Mining Association Rules

Data mining is the process of abstracting unaware, potential and usefulinformation and knowledge from plentiful, incomplete, noisy, fuzzy and stochasticdata. It is a new subject that involves a lot of subjects and develops with thesesubjects. Association rule mining is an important branch of data mining to discoverpreviously unknown, interesting relationships among attributes from largedatabases.This thesis analysis interdependency mining theory and the traditional miningalgorithms. Based on various traditional mining algorisms, this thesis advancesRun-Length Encoding mining algorithms and puts the Run-Length Encodingmining technology into practice in interdependency rule data mining.This thesis encodes Access data into a small amount of data, and then minesthe encoding data in memory directly. The method mentioned above empowers datamining the ability of capturing the inner rule of dynamic data. The merit of thisalgorithms lies in that it needn’t read the information in database repeatedly whilethe data shifts rapidly and can fresh the data in order to enhance the executive speedand improve the effectiveness of dealing with data.The association rule algorithm overcomes the shortcomings of traditionalalgorithms’ timing data dealing. Exactly speaking, it is based on the improvementof Apriori algorithms. As the support is very low, it is possible to produce a lot offrequent set of items. In this case, this association rule algorithm is not the ideal one,but it is not discussed in this thesis.