Font Size: a A A

Fuzzy Association Rules Mining And Application In Industrial Data

Posted on:2015-08-03Degree:MasterType:Thesis
Country:ChinaCandidate:J M GuoFull Text:PDF
GTID:2298330431995534Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
With the rapid development of science and technology and intelligent controltechnology, A lot of historical data accumulated in the industrial production. As oneof the current research emphasis in the field of data mining, excavating potentialdomain knowledge and associated relationship from these historical data can providedecision-making advice and enrich the field experience in the management of industry.The actual process data involve the typical features of highly complexity, insist ofnumerical data and redundancy data. therefore it is difficult to establish accuratemathematical model. Through the traditional association rules mining method willcause the rigid boundary of numeric data, which brings the problems such as potentialinformation loss. In this paper, the fuzzy association rules mining algorithm wasapplied to mining the underlying relationship in the data of industrial managementsystem, providing effective basis for decision-making and scientific plan.Based on the comparison of the current commonly used methods, we firstintroduced the association rules and basic extraction method of fuzzy association rules,and an improved fuzzy association rules mining method based on clusteringoptimization was proposed considering the lack of current fuzzy association rulesextraction algorithm.finally, in this paper, the algorithm was applied to mining thepotential rules in the industrial management system.Aiming at the problem of rigid boundary of numeric data in the traditional miningmethod, in this paper, the algorithm first uses the clustering optimized c-meansclustering algorithm to deal with the original data sets and eliminate the redundantdata, and generate fuzzy partition. For the traditional c-means clustering algorithm issensitive to the initial clustering center, easy to fall into local minimum in the iterativeprocess, and using fuzzy partition to map the original data sets to fuzzy interval. Theexperiment shows that this method can effectively improve the clustering accuracyand improve the mining efficiency, and enhance the semantic intuitiveness andcredibility of association rules.Finally,in this paper, Fuzzy association rules mining algorithm was applied inthe data set of industrial management of energy consumption. Through the design of the variables association rules extraction model, we extract the relevant rules in thedatasets of the management system. And then we put forwarded some optimizationsuggestions.So that it can provide the basis for scientific decision in industrialprocess control field.
Keywords/Search Tags:Association rule extraction, fuzzy clustering, industrial data
PDF Full Text Request
Related items