Fake Media Detection Based on Natural Language Processing and Blockchain Approaches--with otp

Fake news dissemination through social media has become a serious challenge in the digital era, influencing public opinion and creating misinformation at large scale. Traditional fake news detection systems mainly rely on machine learning algorithms such as Support Vector Machine and Random Forest, which suffer from limitations related to adaptability, security, and authorization. To overcome these issues, this project proposes a hybrid approach combining Reinforcement Deep Learning and Blockchain technology for secure and intelligent fake media detection. Natural Language Processing techniques such as stop word removal, normalization, stemming, lemmatization, and vectorization are applied to preprocess news content. The extracted numerical vectors are fed into a Reinforcement Learning model, where each news item is treated as a state and the prediction outcome represents an action. The model rewards users for publishing genuine news and penalizes those spreading fake content. Verified real news is then securely stored in a blockchain, ensuring data integrity, decentralization, and immutability through cryptographic hash verification and consensus mechanisms such as Proof of Work or Proof of Authority. The proposed system improves detection accuracy, enhances security, and ensures reliable news storage using decentralized infrastructure. The BUZZ News dataset is used to train and evaluate the system, demonstrating effective fake news identification and tamper-proof data management.

Keywords: Fake News Detection, Reinforcement Learning, Blockchain, Natural Language Processing, Social Media Security