N.Vamsi, B.Bhagya Lakshmi 03 06 The exponential growth of online content and social media platforms has accelerated the spread of fake news posing serious threats to societal stability democratic processes and public trust Automatic fake news detection has thus emerged as a crucial research problem in the domains of Natural Language Processing NLP and Machine Learning ML This paper proposes a FeatureBased Optimized Multiclass Support Vector Machine MSVM classification framework to effectively detect and classify fake news The proposed system leverages advanced text preprocessing techniques hybrid feature extraction methods including TFIDF ngrams sentiment analysis and linguistic cues and feature optimization techniques such as Chisquare and Principal Component Analysis PCA The optimized MSVM classifier is designed to handle highdimensional feature spaces and multiclass classification efficiently Experimental evaluation demonstrates that the proposed system significantly improves classification accuracy robustness and scalability compared to traditional methods The system is suitable for realtime deployment in social media monitoring and news verification platforms Copyright (c) Journal of Science Engineering Technology and Management Science. All rights reserved 2026