Detecting Fake News Through Probabilistic Sentiment Model and Sentence Embedding Technique

Abstract

False information is purposely disseminated through fake news, aiming to spread hoaxes, propagation, and disinformation. Analyzing news truthfulness is a challenging problem because there is a massive information on social networking sites which turns out to be very difficult to manually analyze. Moreover, the impact of fake news is tremendously huge for internet users. As a result, academic research related to filtering and banning fake news has been highly demanding in very recent years. The paper contributes to this field by using a probabilistic sentiment model with a sentence embedding method, addressing the limitations of approaches relying on TF-IDF or Word2Vec methods. The proposed method incorporates probabilistic sentiment scores and sentence embedding features to improve the accuracy of the fake news detection system, employing Naïve Bayes and Support Vector Machine classifiers. The proposed method achieves a high accuracy of 99% F1-score by combining these features.