A multiagent system (MAS) for the generation of initial centroids for k-means clustering data mining algorithm based on actual sample datapoints

Clustering is a technique in data mining to find interesting patterns in a given dataset. A large dataset is grouped into clusters of smaller sets of similar data using k-means algorithm. Initial centroids are required as input parameters when using k-means clustering algorithm. There are different methods to choose initial centroids, from actual sample datapoints of a dataset. These methods are often implemented through intelligent agents, as the later are very commonly used in distributed networks given that they are not cumbersome for the network traffic. More over, they overcome network latency, operate in heterogeneous environment and possess fault-tolerant behavior. A multiagent system (MAS) is proposed in this research paper for the generation of initial centroids using actual sample datapoints. This multiagent system comprises four agents of k-means clustering algorithm using different methods namely Range, Random number, Outlier and Inlier for the generation of initial centroids.