AN OPTIMAL UNSUPERVISED TEXT DATA SEGMENTATION USING GENETIC ALGORITHM

Abstract

The popularity of information available in electronic forms has been rapidly growing in the last decade, and turn into a golden mount containing extremely unstructured data for the researchers. Extracting interesting information and knowledge from such data creates promising future path into the era of text mining. The roots of text mining lie in most related research areas clustering, classification, information retrieval, machine learning and soft computing paradigms. Among all, Clustering is an unsupervised methodology having the ability to form meaningful natural groups of objects from given unlabeled data. A large number of clustering algorithms based on K-Means have been proposed on variety of domains for different types of applications none of these algorithms is suitable for all kinds of applications. This motivated and find a room for new clustering algorithm that is more efficient and optimal with computationally feasible. Genetic algorithms are randomized optimization techniques guided by the principles of evolution and natural genetics, having a large amount of implicit parallelism. To improve the text data segmentation accuracy the authors proposed An Optimal Unsupervised Text Data Segmentation Model using Genetic Algorithm so called OUTDSM. The encoding strategy, fitness function and operators of proposed OUTDSM works together and achieve high accuracy rated optimal clusters. Additionally, the nature of biological diversity of OUTDSM prevents the population from stagnating at any local optima and promises to arrive at global optima. The experimental results proving this claim are given in this paper.