Huddle Based Harmonic K Means Clustering Using Iterative Relocation on Water Treatment Plant Data Attributes

Clustering is one of the important methods in data mining to weight the distance between the cluster objects effectively. Existing K-means algorithm with Affinity Propagation (AP) algorithm captures the structural information of texts, and improve the semi supervised clustering process. Seeds Affinity Propagation (SAP) provides the detailed distance measurement but it takes the longer execution time to perform the clustering process. Weighted Clustering Ensemble (WCE) algorithm on the existing work provides an effectual technique on the temporal data clustering. WCE are not effective in developing the speed clustering on the data mining system. To reduce the execution time on clustering of the large capacity of dataset, Huddle based Harmonic K-means Clustering (HH K-means Clustering) mechanism is proposed in this paper. HH Kmeans Clustering first performs the initial screening process to divide them into a cluster and non-cluster data form. Secondly, HH K-means Clustering develops the Iterative Relocation (IR) technique to improve the speed of the clustering process. Iterative Relocation (IR) technique calculate and estimates the cluster center objects (i.e.,) Centroids in HH K-means Clustering. The IR technique performs the initialization, relocation of objects and cluster updating process to reduce the computation of distance measure on the every data objects. Huddle groups the data objects using the Harmonic K-means Clustering and calculate the clustering time and clustering rate. So, HH K-means clustering add some improved conditions and factors to enhance the effectiveness of data mining clustering system with approximately 7 % improved cluster rate. Experiment is conducted using Water Treatment Plant Data Set on the factors such as execution time, huddle based clustering rate and entropy level.