Sentiment analysis techniques, challenges, and opportunities: Urdu language-based analytical study.

Area of Science:

• Natural Language Processing
• Computational Linguistics

Background:

• Sentiment analysis is effective for high-resource languages but limited for low-resource languages like Urdu due to scarce textual data.
• Urdu, spoken by over 169 million people, lacks sufficient corpora, language parsers, and pre-trained models for effective sentiment analysis.
• Existing research highlights shortcomings in Urdu sentiment analysis, impacting performance and applicability.

Purpose of the Study:

• To systematically explore and evaluate machine learning-based Urdu sentiment analysis studies.
• To identify key challenges and limitations in the current literature.
• To propose research objectives and questions for advancing Urdu sentiment analysis.

Main Methods:

• Systematic literature review of Urdu sentiment analysis.
• Searched and filtered digital repositories for relevant studies.
• Inspected forty selected articles and extracted relevant data.

Main Results:

• Identified limitations including small corpora, inadequate language parsers, and a lack of pre-trained models.
• Sentiment classification performance can be enhanced by addressing word sense disambiguation and utilizing massive datasets.
• Urdu language constructs, context-level analysis, pre-processing, and lexical resources require improvement.

Conclusions:

• Overcoming resource limitations is crucial for improving Urdu sentiment analysis.
• Further research is needed in areas like language parsing, pre-trained models, and handling language-specific constructs.
• Advancements in these areas will enhance the performance and utility of Urdu sentiment analysis.