Comparative Analysis of Nature Inspired Algorithms on Data Clustering

K-Means clustering is well accepted clustering algorithm that huddle similar data objects in a simple and quick way. The convergence speed of K-Means clustering is quite appreciable but it has drawback of getting stuck into local optima. Hence, optimal clustering results are not attained. Nature inspired algorithm when integrated with clustering algorithm provides global optimal solution. The paper analyzes three nature inspired algorithms i.e. firefly algorithm, bat algorithm, and flower pollination algorithm integrated with K-Means clustering. The study is performed on four real life datasets obtained from UCI machine learning repository and two simulated datasets. Algorithms are evaluated on the basis of number of fitness function and CPU time per run. It is observed from experimental study that integrated flower pollination algorithm with K-Means overrule the other two algorithm on each datasets.