Parallel processing of enhanced K-means using OpenMP
D. S. Bhupal Naik, Sunil Kumar, Sri Ramakrishna · 2013
Cluster Analysis plays a vital responsibility in scientific investigation and business applications. K-Means clustering algorithm is broadly used as a partitioning technique. K-Means clustering algorithm is not much suitable for huge voluminous of data sets. Iterative clustering with K-Means has more Execution time. To avoid such, A Parallel Partitioning of enhanced K-Means algorithm using OpenMP is proposed to handle the outliers with optimized execution time without affecting the accuracy. The experiments are performed on diabetes, soya beans and supermarket by considering multi-core systems with 768, 683 and 4627 instances respectively. The proposed method outperforms with an accuracy of 74.76% for diabetes dataset with an execution time of 56secs, soya beans datasets with an accuracy of 82.34% with an execution time 54secs and supermarket datasets with an accuracy of 80.45% with 54secs of execution time.