Design of an Efficient Framework to Enhance the Clustering Performance in Data Mining

Abstract

Data mining method is generally used for determining more important newlineinformation in an enormous dataset. The mining of data is a procedure of being newlineacquainted with consistent patterns in a huge dimension of data applied to the newlinetechniques of unsupervised clustering, statistics, genetics, and radial basis newlinefunction. Data mining concepts extract good information obtained from the newlinedataset where those particular datasets are created well in clusters shape with newlineconvergence. The clustering techniques can be categorized namely as partitioning newlineclustering, hierarchical clustering, density-based clustering, and grid-based newlineclustering. It has more utilities to carry the data mining as an essential part of the newlinebusiness. newlineWe have proposed an efficient framework of clustering approach, and its method newlinethat improves clustering metrics, analyze the clustering of k-means with other newlineapproaches using the software tools, propose and analysis the fitness objective newlinefunction using Genetic Algorithm (GA), and analysis the clustering metrics SSE newlineusing Radial Basis Function of Neural Network of ANN. newlineAn efficient framework is being presented for producing the good quality of newlineclusters. Evaluate metrics related to performance with the component of clusters. newlineHere, we have discussed the every component of the framework. The component newlineof the framework consists are first components proposed methodology, and newlineproposed algorithm is hybridized via PCA and PSO; second component is a newlinestatistical analysis with software tools; Third component utilizes the Genetic newlineAlgorithm(GA), and fourth component RBFN of ANN theory, using datasets. newlineThere is the first component discussed to the proposed algorithm of clustering newlineand it is implemented on various sizes of the datasets. We have implemented newlinean experiment on MATLAB R2013a to measure the metrics of the cluster and newlinealso measure the fitness of fitness function values using particle swarm newlineoptimization has been accomplished through the critical literature survey, collects newlinethe ideas of experts