FAKE NEWS DETECTION SYSTEM ON INSTAGRAM USING MACHINE LEARNING MULTI-MODEL METHOD

Abstract

The consumption of news on social media is growing in popularity. Social media appeals to users because it is low-cost, user-friendly, and rapidly disseminates information. False information does, however, also circulate, in part because to social media. It's getting harder and harder to ignore fake news because of the harm it causes to society. However, depending only on news content usually leads to poor detection effectiveness because fake news is designed to look legitimate. As such, a detailed comprehension of the relationship between social media user profiles and fake news is necessary. This study examines the use of machine learning algorithms to detect fake news. It covers significant subjects like user profiles, dataset analysis, and feature integration. The study integrates attributes to provide large feature sets. When dealing with high-dimensional datasets, Principal Component Analysis (PCA) is a helpful technique for dimensionality reduction. The study uses datasets from "Instagram," which include a variety of data processing techniques, to extensively analyze several machine learning models. The evaluation of the Random Forest classification model is further improved via curve analysis. The outcomes show the best feature and model pairings, with our model outperforming the competition.