H-MFN: A HYBRID MULTIMODAL DEEP LEARNING MODEL FOR FAKE NEWS DETECTION IN INDIAN NEWS AND SOCIAL MEDIA
Arvind Sharma1, Pratibha Soni2, Hukam Chand Saini3
1M.Tech. Scholar, 2Assistant Professor, 3Associate professor
Department of Computer Science, Jagan Nath University, Jaipur
Abstract: In today’s digital world, fake news spreads very fast through social media and news websites. This fake information can confuse people and lead to wrong decisions. So, it is very important to build a system that can detect fake news in a quick and accurate way. In this study, we focused on fake news detection using simple and advanced methods. We collected both real and fake news data from different platforms such as Facebook, X (Twitter), Instagram, and news websites. The data includes different topics like politics, education, technology, and entertainment. We used machine learning and deep learning models to understand and detect fake news. To make the data more useful, we cleaned it and added labels like “real” or “fake” with the help of trained annotators. We also made sure that the labels were reliable by checking agreement scores using special statistical methods. The models were trained using this labelled data, and their performance was checked using accuracy, precision, recall, and F1 score. We also studied the news headlines and descriptions across different categories. We looked at total words, unique words, and headline length. This helped us understand how fake and real news are written differently. Our final system showed good performance in detecting fake news in the English language. Current research will help in building better tools to identify fake news in English. It can also support journalists, readers, and fact-checkers to understand which news is true and which is not. In the future, we aim to improve this system by adding more news types, including images and videos, and using even smarter models. This study is an important step towards reducing the harmful effects of fake news in society.
Keywords: Fake News Detection, Stacked Ensemble Learning, Text Classification, BERT, LightGBM, Support Vector Machine (SVM), TF-IDF, XGBoost, Hybrid Machine Learning Model
VOLUME 10 ISSUE 02 2026: 24 – 49