WCSE

Comparative Study of Fake News Detection Using Machine Learning and Neural Network Approaches

Abstract— In today’s era, fake news related to politic, finance, reputation, health and education gets create and spread like wild fire within the people as the improvement of using social media. Spreading of misinformation is a cause of great concern for human society. Detecting fake news on social media becomes a challenging problem which turns out to be very difficult to manually analyze because more and more online news is increasing on social network. Although a lot of fake news detection researches have shown some significant results and improvements by using different classification algorithms and feature extraction methods, it still has some gaps to meet the important necessities in classifying news. To address this problem, this paper investigates a fake news detection model using machine learning and neural network approaches with frequency-based and word-embedding feature extraction methods. The system performs the experiments upon “ISOT” fake news benchmark dataset. The experimental results show that exploitation of Long Short Term Memory (LSTM) with Glove feature vector achieves better classified results than Feed Forward Neural Network (FFNN), Naïve Bayes (NB) and Support Vector Machine (SVM) in classifying news.

Index Terms— misinformation, social media, fake news detection, machine learning, neural network approach

May Me Me Hlaing, Nang Saing Moon Kham
University of Computer Studies, Yangon, MYANMAR

[Download]

Cite: May Me Me Hlaing, Nang Saing Moon Kham, "Comparative Study of Fake News Detection Using Machine Learning and Neural Network Approaches, " Proceedings of 2021 the 11th International Workshop on Computer Science and Engineering (WCSE 2021), pp. 59-64, February 25-27, 2021.