A New Hybrid Approach for Data Clustering Analysis using Hybrid Fuzzy C-Means and Fuzzy Particle Swarm Optimization

Abstract

The most reliable and robust document clustering algorithms have played a significant role in document clustering for active celestial navigation, summarize, and organization of data. K-means, a common partitioned clustering algorithm, has the potential to fix powerful mathematical tasks in a short period of time.  In this paper, an effective document clustering process is performed using a new hybridized k-means based genetic algorithm. Thereby, this new hybridized algorithm has prevented the solutions trapping inside the local optima. By the fact, the better solutions generated using this hybridized algorithm tackles clustering issues during cluster formation with a huge number of documents/data. The hybrid algorithms once equated with traditional methods on two benchmark text document datasets afford improved superiority document clusters in terms of two standard document clustering assessment measures Entropy and F-Measure. Experimental validation was conducted on four data set categories. Experimental outcomes have proved the effectiveness of proposed hybridized algorithms compared to other conventional optimization algorithms such as, PSO and K-means in terms of clustering efficiency, accuracy, and F-measure.