Fake News Detection Using Machine Learning Techniques

Project Overview

The emergence of the World Wide Web and the swift uptake of social media platforms like Facebook and Twitter have opened up avenues for the dissemination of information that were previously unprecedented in human history. The current widespread use of social media platforms has resulted in a surge of information creation and sharing by users, surpassing any levels seen before. However, a significant proportion of this information can be misleading and bears no connection to reality. The wide spreading of fake news is a matter of serious concern due to its ability to cause a lot of social and national damage with destructive impacts. To overcome the impact of fake news, this research focuses on leveraging the power of machine learning to classify news into Fake or Real.
Four Independent machine learning algorithms were used for news classification, also, an ensemble method was used to strengthen the prediction accuracy. The four algorithms are:

• 1. Logistic Regression
• 2. Decision Tree
• 3. Naïve Bayes
• 4. Support Vector Machine
• 5. While ensemble Random Forest was used

Out of the four algorithms used, Support Vector Machine recorded the highest accuracy of 93%, followed by Logistic Regression at 91%, followed by Naïve Bayes at 84%, while the Decision Tree recorded the least accuracy at 80%`. The Ensemble Learning method recorded 93% accuracy, which outperformed all the used algorithms except Support Vector Machine.