Novel Sentiment Majority Voting Classifier and Transfer Learning-Based Feature Engineering for Sentiment Analysis of Deepfake Tweets

This research presents a novel approach to sentiment analysis of deepfake tweets by developing a sentiment majority voting classifier combined with transfer learning-based feature engineering. The objective is to accurately distinguish between human-generated and robot-generated tweets, leveraging the power of machine learning and natural language processing. A two-column dataset containing the tweet text and corresponding labels (human or robot) serves as the input. We employ Bidirectional Encoder Representations from Transformers (BERT) for feature extraction, taking advantage of its superior contextual understanding of text. These features are then classified using a Random Forest algorithm, known for its robustness and effectiveness in handling large datasets. The proposed model integrates the strengths of both BERT and Random Forest, providing an efficient and scalable solution for deepfake tweet detection. This approach contributes to enhanced detection accuracy, offering a practical framework for identifying automated content in social media environments.

Keywords: BERT and Random Forest, Kaggle dataset.