The Future of Media with AI Technologies: Detecting Fake News

Abstract

In this study, we review the important role played by artificial intelligence and machine learning technologies in addressing the problems and challenges of detecting fake news.

With the advancement of social media, detecting fake news has become a challenging issue. Fake news can be spread, leading to negative consequences for society and serious challenges.

We faced challenges in this study, the most important of which was obtaining a good and comprehensive dataset for the topic. The data available online was very large, requiring a lot of time and effort to classify and process it. To overcome this difficulty, we used natural language processing to process the data by cleaning the data, reducing its dimensions, and removing duplicate values.

The aim of the study was to evaluate three machine learning algorithms and their effectiveness in classifying fake and real news on a dataset containing 23,503 news samples. We divided the data into a test and training set with a ratio of 75% for training and 25% for real news, maintaining a class balance between the two categories.

The results outperformed the logistic regression algorithm with a performance of 98% accuracy, followed by SVM with 95%, while Naïve Bayes achieved the lowest result with 86%.