Devising a Machine Learning-Based Instagram Fake News Detection System Using Content and Context Features

Fake News is known as one of the most serious threats to our current society and when it combines with democracy, it can result in disastrous consequences. People are not good at detecting fake news because of some cognitive distortions like confirmation bias and naive realism. Lately, AI has been considered an intelligent assistant for humans in detecting fake information. In this paper, we propose an automated fake news detection system tailored for Instagram. Based on the unique characteristics of Instagram, some content and context features have been generated and fed into our classifying module. Different Machine Learning algorithms like Support Vector Machines (SVM), Logistic Regression, Naive Bayes, Random Forest, and K-Nearest Neighbors (KNN) have been applied in the proposed system. Also, an Instagram fake news dataset has been created for experiments. The results show that our system can classify Instagram fake news with high precision while KNN is the most powerful method with 99 percent of F1-measure. Some of our proposed features like the number of posts, mentioning URL in bio, the average number of comments per post, engagement rate, the average number of likes, and the number of fake followers are among the most important features in various ML algorithms.