Higher Classification of Fake Political News Using Decision Tree Algorithm over Random Forest Algorithm

Abstract

The current project aims to model and compare the performance of fake news detectors using machine learning algorithms to recognize fake news connected to political topics with high accuracy. The Decision Tree algorithm and the Random Forest algorithm are two algorithms. The methods were developed and evaluated on a dataset including 44,000 samples. Implemented each algorithm through programs and performed ten iterations with different scales of false feeds and factual feeds classification were identified. The G-power test is around 80% accurate. For detecting false political news, the Decision Tree algorithm had a mean accuracy of 99.6990, and the Random Forest approach had a mean accuracy of 98.6380, according to the trial results. The significance of accuracy is p=0.001, indicating the efficacy of the classifier. This research aims to use a novel strategy for contemporary Machine Learning Classifiers to predict fake political news. The comparison results reveal that the Decision Tree method outperforms the Random Forest technique.