Performance Enhancement of Data Mining Algorithm for High Dimension Iot Dataset

Abstract

The procedure of data mining is taking the raw data and finding the best possible results among the data which make sense and can be used for the betterment.K-means is now the most often used algorithm in data mining, as well as one of the earliest algorithms for all of this objective. The technique is modified throughout time to meet the demands of the researchers in order to get the desired outcomes, but the problems with the classical k-means remain there.

Therefore to overcome the problem of initial centroid selection is been considered by David Arthur and an enhanced k-means++ algorithm is been developed which chooses the centroid as per the strategy developed by David Arthur.  In the proposed paper, we examine the Fusion K-Means algorithm over the normal K-Means algorithm to see the changes in the result obtained by both the implementation. The major area of work done in this paper is choosing the initial centroids so that they can handle the high dimensionality data set. Sothat the more accurate and optimized data clusters are carried out