Evaluation and Efficient Initial Centroid Selection of New Algorithm for High Dimensional Data?

Clustering is a division of data into groups of similar objects. Each group called cluster, consist of objects that are similar between themselves and dissimilar compared to the object of other groups. Dimensionality reduction is the transformation of high-dimensional data into meaningful representation of reduced dimensionality that corresponds to the intrinsic dimensionality of the data. This paper proposed a new algorithm of data partitioning based k-means will perform the data axis with highest variance will be chosen as the principal axis for data partitioning. The data partitioning tries to divide data space into small cells or clusters where inter cluster distance are large and intra cluster distance are small as possible. The experimental results show that the proposed algorithm mainly focuses on reducing the Sum of the Squared Error and increasing the clustering accuracy than the existing algorithm of k-means algorithm and Pam algorithm.