Thai Fake News Detection Based on Information Retrieval, Natural Language Processing and Machine Learning

Area of Science:

• Computational Linguistics
• Artificial Intelligence
• Information Science

Background:

• Fake news poses a significant societal challenge, necessitating effective detection and mitigation strategies.
• The dynamic nature of fake news complicates accurate identification and prevention of its spread.
• Developing robust systems for fake news detection is crucial for maintaining information integrity.

Discussion:

• A novel framework for Thai fake news detection integrates information retrieval, natural language processing (NLP), and machine learning (ML).
• The research involved a two-phase approach: data collection via web crawling and feature extraction using NLP techniques.
• A comparative analysis of various ML classification models was conducted to identify the most effective approach.

Key Insights:

• The Long Short-Term Memory (LSTM) model significantly outperformed other evaluated classification algorithms (Naïve Bayesian, Logistic Regression, K-NN, MLP, SVM, Decision Tree, Random Forest, Rule-Based Classifier).
• Feature extraction using NLP techniques proved vital for enhancing the performance of machine learning models in fake news detection.
• The study successfully deployed an automatic online fake news detection web application.

Outlook:

• Future research could explore advanced deep learning architectures for improved fake news detection accuracy.
• Expanding the framework to encompass multiple languages and diverse data sources will enhance its generalizability.
• Continuous monitoring and adaptation of detection models are essential to counter evolving fake news tactics.