An optimized clustering technique for higher dimensional data

Abstract

Data mining in databases is the automatic extraction of implicit and interesting patterns from large data collections. Data mining is a field at the intersection of computer science and statistics and is the process that attempts newlineto discover patterns in large data sets. It utilizes methods at the intersection of newlineartificial intelligence, machine learning, statistics and database systems. newlineClustering is a technique in data mining which deals with huge amount of newlinedata. Clustering is intended to help a user in discovering and understanding newlinethe natural structure in a data set and abstract the meaning of large dataset. It newlineis the task of partitioning objects of a data set into distinct groups such that newlinetwo objects from one cluster are similar to each other, whereas two objects newlinefrom distinct clusters are dissimilar. newlineMany situations have been adapted in the issue of clustering high newlinedimensional data. By computation of similarity measure between the data newlinepoints of a different class of data points while clustering data points with a newlinesmall size can be done in an easier way. As the size of dimension becomes newlinebigger, the similarity measure between data point also becomes difficult and newlinethus the issues also becomes tougher. A false indexing ratio is introduces and newlinewith the multiple class names, same data point may be assigned, where newlinecomputation of similarity between data points should be done considering all newlinethe dimensional values. In this work, proposed the Fuzzy C Means (FCM) newlineclustering, Fast clustering-based feature Selection algorithm (FAST) and newlinehubness in clustering for high dimensional data. By the usage of graphtheoretic newlineclustering, the features are spit into clusters. newline newline