A systematic literature review on meta-heuristic based feature selection techniques for text classification.

Area of Science:

• Data Science
• Machine Learning
• Natural Language Processing

Background:

• Feature selection (FS) is crucial for optimizing data science applications, particularly text classification.
• Excessive and irrelevant features negatively impact classifier performance.
• Traditional and meta-heuristic (MH) techniques are employed for FS.

Purpose of the Study:

• To systematically analyze MH techniques for FS in text classification from 2015-2022.
• To identify and evaluate the strengths and weaknesses of various MH techniques.
• To compare MH techniques against traditional FS methods.

Main Methods:

• Conducted a systematic literature review of 108 primary studies.
• Focused on studies published between 2015 and 2022.
• Searched databases including Scopus, Science Direct, and Google Scholar.

Main Results:

• MH techniques demonstrated superior performance compared to traditional FS methods.
• Identified key MH techniques and their associated advantages and disadvantages.
• Highlighted the efficiency and effectiveness of MH approaches for feature selection.

Conclusions:

• MH techniques are highly effective for feature selection in text classification.
• Further research into MH techniques, like Ringed Seal Search (RSS), can enhance FS capabilities.
• MH techniques offer significant potential for improving various data science applications.