Unsupervised Text Feature Selection Technique Based on Particle Swarm Optimization Algorithm for Improving the Text Clustering

Abstract

After incensing the amount of text information on internet web pages, the dealing with this information is very complex due to the volume of information. Text clustering technique is an appropriate task to deal with a huge amount of text documents by grouping set of documents into groups. Text documents contain uninformative features, which decrease the performance of the text clustering technique. Feature selection is an unsupervised technique used to select informative features by creating a new subset of informative features. This technique used to improve the performance of the underlying algorithm. Latterly, several complex optimization problems are success solved by metaheuristic al-gorithms. In this paper, we proposed the Particle swarm optimization algorithm to solve the feature selection problem, namely, (FSPSOTC). The feature selection technique encourages the k-mean text clustering technique to obtain more accurate clusters. Experiments were conducted using four standard benchmark text datasets with different characteris-tics. Experimental results showed that the proposed method (FSPSOTC) is enhanced the performance of the text clustering technique by dealing with a new subset of informative features.

Advances in Meta-Heuristic Optimization Algorithms in Big Data Text Clustering

“…Abualigah et al [96] proposed a new technique for solving the feature selection problem based on the HS algorithm to search for the best subset of informative features. Then, the proposed method is utilized for clustering texts and is summarized as FSHSTC, FS technique using HS algorithm for the TC technique, where it can overcome the other methods drawbacks in improving the performance of the text clustering.…”

Approach to text data clustering based on molecular chemical reactions in agricultural and food industries

“…Text data clustering is one of the problems of information retrieval [1,2,3].Text data clustering is used to group unordered text documents [1,4]. The purpose of cluster analysis of text data is to detect groups of semantically similar documents among a collection of given text data without predefined categories (characteristics) of grouping.…”

“…In the literature, there are many scientific approaches to text data clustering [3,5,6,7,8,9,10]. The k-means local search algorithm is most widely used [8].…”

Optimization Algorithms to Solve Feature Selection Problem: A Review

“…Embedded methods works with linear classifiers such as SVM are embedded in the algorithm as expanded functionality. It is also able to capture dependencies at a lower computational cost than other methods [1], [38], [50], [51]. This paper is organized as follows: A general description of the feature selection problems is presented in Section 2.…”