Fake news detection in social media based on sentiment analysis using classifier techniques

Area of Science:

• Natural Language Processing
• Machine Learning
• Computational Social Science

Background:

• Fake news proliferation on social media poses significant societal challenges.
• Accurate fake news detection is crucial for maintaining information integrity.
• Existing methods often struggle with the nuances of online content and missing data.

Discussion:

• The study proposes a novel fake news detection framework incorporating sentiment analysis.
• Key features are extracted using sentiment propensity scores, Term Frequency-Inverse Document Frequency (TF-IDF), and a lexicon-based scoring algorithm.
• A multiple imputation strategy (MICE) is employed to effectively handle missing variables in social media datasets.

Key Insights:

• Sentiment analysis significantly enhances fake news detection accuracy.
• The integration of TF-IDF and sentiment features provides robust content analysis.
• The proposed method achieved a remarkable 99.8% accuracy in classifying various truthfulness levels of statements.

Outlook:

• This research offers a highly accurate and efficient solution for combating online misinformation.
• The methodology can be adapted for diverse applications in content moderation and digital forensics.
• Further research could explore real-time detection and cross-platform analysis of fake news.