3 WAY SHIELD ENHANCED K - MEAN NORMALISATION BY INTELLIGENT INIITIALISATION OF CENTROID

Learning to learn" is a notion of full excitement in advance research within machine learning .K - means is a most widely used approach in unsupervised machine learning algorithms though it was proposed 50 years ago,but undoubtedly is usually very fast,but its efficiency is highly depends on the placement of initial centroids.Thus,the proposed study is based on previously made the effort and improvements over traditional K - mean clustering scheme by introducing the impact of one the best initialization method named as PCA - PART,inputting normalized linear transform geometric data - sets over 2D hyper plane.This amalgamation in k - means clustering not only give provision for noise found to be the most issue factor in centroid based clustering but also improve efficiency of k - means by airing the data - set and removing hurdles leading to fine - grained initial centroids.