An Efficient Partitional Clustering Algorithm Based on Splitting Merging Strategy

摘要

The A-means has been one of the most accepted partitioning clustering algorithms among a wide range of clustering algorithms, due to its superior scalability and efficiency. But, key drawbacks of Ameans algorithm is that it usually creates empty clusters depending on initial cluster centers and number of clusters dependency. In order to tackle these challenges, we have proposed an efficient partitional clustering algorithm utilizing A-means clustering. The concept proposed in the work can generate the initial clusters to initialize A-means clustering. The initial cluster centers of the -means clustering are identified by applying the splitting and merging strategy, which employs density and mean parameter of the initial partitions. The experimentation is carried out on wine datasets and the experimental results ensured that the proposed algorithm has achieved better clustering accuracy and less computation time compared with the k-means clustering algorithm.