Hierarchical Parallel Processing for Data Clustering in GPU Using Deep Nearest Neighbor Searching

Abstract

Today growing number of corporations and research groups can rely on this new tool for are hitching up artificial intelligence horsepower. Insurance, Banking, Retail, Telecom and many other such sectors can find it fruitful for optimizing their options. artificial intelligence applications are becoming more prevalent: the improving tax collections and detecting tax fraud; improving its health care for employees while reducing the corporation’s costs; We are beginning to figure out how to mine these growing mountains of artificial intelligence, data, and parallelism makes the mining operations possible. We need to provide efficient solution to solve data clustering in GPU. Hierarchical parallel processing method is applied to find the data clusters in GPU. Deep nearest neighbor searching algorithm is used to create deep belief network and predict the accuracy. The efficiency is determined in the training set using the mean square error rate. The obtained results are compared with the traditional techniques. The result is tested by using TensorFlow using different GPU time slots.