Client Based Power Iteration Clustering Algorithm to Reduce Dimensionality in Big Data

Clustering is a group of objects that are similar among themselves but dissimilar to objects in other clusters. Clustering large dataset is a challenging task and the need for increase in scalability and performance formulates it to use parallelism. Though the use of Big Data has become very essential, analyzing it is demanding. This paper presents the (pC-PIC) parallel Client based Power Iteration clustering algorithm based on parallel PIC originated from PIC (Power Iteration Clustering). PIC performs clustering by embedding data points in a low dimensional data derived from the similarity matrix. In this paper we have proposed a client based algorithm pC-PIC that out performs the job done by the server and reduces its execution time. The experimental results show that pC-PIC can perform well for big data. It’s fast and scalable. The result also shows that the accuracy in producing the clusters is almost similar to the original algorithm. Hence the results produced by pC-PIC are fast, scalable and accurate.