Fake news detection with machine learning methods

S. Aphiwongsophon, and P. Chongstitvatana

Abstract

Fake news detection is an interesting topic for computer scientists and social science. Because of the recent growth of the online social media fake news has great impact to the society. There is much information from disparate sources among various users around the world. Twitter is one of the most popular applications that are able to deliver appealing data in timely manner. Developing a technique that can detect fake news from Twitter is becoming a necessary and challenging task. This article proposes a machine learning method which can identify fake news from Twitter data. The experiment is carried out with three widely used machine learning methods: Naïve Bayes, Neural Network and Support Vector Machine using the Twitter data collected from October to November 2017 on two particular topics in Thailand. The results show that all three methods can detect fake news in this data set accurately. The accuracy of Naïve Bayes method is 96.08 percent, Neural Network 99.89 percent and Support Vector Machine 99.89 percent. Furthermore, we analyse the data of fake news and point out some of its characteristics.

link to data set (.csv file, 17M)

explanation of data format

last update 13 Feb 2019