Enhancement of random forest algorithm applied in fake news detection

Abstract

The widespread increase of fake news presents a serious obstacle in today's information-sharing situation, making it more difficult to distinguish fact from fabricated information. The study aims to improve machine learning algorithms' capacity to identify fake news to address this significant issue. To improve detection accuracy, Researchers applied a method that combines the Random Forest algorithm with Adaptive Boosting, or AdaBoost. By utilizing ensemble learning techniques, our method successfully leverages the collective intelligence of numerous decision trees, outperforming traditional approaches. By carefully experimenting with hyperparameters using the Random Search, The proposed method improved the algorithm's capacity to identify significant patterns suggestive of false information. The integration of AdaBoost to Random Forest produces a significant improvement, from 92.56% to 99.79% accuracy average difference, with an accuracy gain of 7% over the baseline Random Forest model. This improvement emphasizes how effective ensemble methods are at negotiating the complexities in terms of fake news detection. This study adds to the ongoing efforts to protect the integrity of information in the digital era by showcasing the efficacy of sophisticated machine-learning approaches in addressing the complex problem of fake news. The study underscores the critical importance of continuous innovation and adaptability in combating the increase of misinformation. Through interdisciplinary collaboration and technological advancements, the aim is to fortify defenses against the detrimental spread of false information, fostering an information ecosystem that prioritizes truth and resilience.