Indian Journal of Science and Technology

Abstract

Objectives: Extracting relevant information from large database is attaining huge significance. Clustering of relevant information from large database becomes difficult. The major objective of this work is to proposed novel clustering methods for solving clustering problem. Methods/Statistical Analysis: This proposed work introduces possibility of a novel approach Hybrid Artificial Bee Colony-Firefly Algorithm (HABC-FA) for clustering to solve the clustering problem in the benchmark datasets like Fisher’s iris dataset. Here this FA incorporates its genome behavior of fireflies to accomplish the optimal clustering solution with ABC. The performance of this novel algorithm Hybrid ABC-FA is then compared with existing clustering algorithms like the ABC and hybrid Particle Swarm Artificial Bee Colony (PSABC) with regard to different statistical criteria making use three different types of benchmark datasets. Findings: The experimentation results prove that the proposed scheme performs better than the existing Swarm Intelligence (SI) based algorithms like ABC and PSABC in terms of speed and success rate and the proposed HABC-FA algorithm performance evaluates by using clustering parameters like recall, precision and F-measure. Application/Improvements: HABC-FA is proposed for the purpose of solving the clustering problem in the benchmark datasets like Fisher’s iris dataset 
Keywords: Artificial Bee Colony Algorithm (ABC), ABC-Particle Swarm Optimisation (PSO) (PSABC), Clustering, Hybrid Artificial Bee Colony with Firefly Algorithm (HABC-FA), Swarm Intelligence (SI)