A novel approach to text clustering using genetic algorithm based on the nearest neighbour heuristic
Debjani Mustafi, Abhijit Mustafi, G. Sahoo
Abstract
In this paper, we propose a novel clustering algorithm which uses a weighted combination of several criteria as its fitness function. We demonstrate the suitability of the new method in the case of clustering text documents. The proposed algorithm leverages the concept of nearest neighbour separation (NNS) to enhance the separation of the clusters and also outlines a heuristic to compute the NNS. A new parameterized fitness function has been proposed which can be tuned to provide more weightage to the traditional metrics based on inter- and intra-cluster distances of clusters or on the NNS. Genetic Algorithm has been used to perform the actual clustering and the results obtained has been compared with the traditional K-Means algorithm. The performance of the algorithm has been tested on different standard datasets, and the results have been presented.