Arabic text classification using hybrid feature selection method using chi-square binary artificial bee colony algorithm
Musab Mustafa Hijazi, Akram M. Zeki, Amelia Ritahani Ismail · The International Islamic University Malaysia Repository (The International Islamic University Malaysia) · 2021
Text classification is a popular method in data mining. It is utilized to get valuable information from the vast quantity of data. Feature selection is a crucial step in Text classification. It is a vital preprocessing technique for powerful data analysis, where only a subset from the original data features is chosen by removing noisy, irrelevant, or redundant features. In this paper, a feature selection method utilizing the combination of chi-square and Artificial Bee Colony (ABC) is proposed. Chi-square, a filter method that is computationally fast, simple and has the ability to deal with a large dimensional feature, is used as the first level of the feature selection process. After that, the wrapper method, Artificial Bee Colony algorithm, is used as the second level where Naive Base is used as a fitness function. The results showed that a reduced number of features outperformed classification accuracy to that using the original features set. Furthermore, the proposed method had a better performance compared with the chi-square method and the ABC algorithm as a feature selection method