Fake News Detection System

Digitalization has transformed the way information is created, shared, and consumed across social media platforms and online news portals. The rapid proliferation of misinformation and fake news has become a critical societal challenge, influencing public opinion, democratic processes, and public health. Small news organizations and individual users frequently lack access to reliable verification tools and depend on manual fact-checking, which is time-consuming and difficult to scale. This paper proposes "Fake News Detection System", a web-based integrated platform designed to automatically classify news articles and social media posts as genuine or fake using machine learning and natural language processing techniques. The proposed architecture separates the presentation, application, and data layers. A React.js web client communicates with a Python Flask REST API, which integrates machine learning models trained on textual features. MySQL provides centralized relational storage for user accounts, news records, prediction history, and feedback data. JWT-based authentication and role-based authorization control access for administrators and registered users. The system provides an administrative dashboard for model monitoring, prediction analytics, and user management. The research follows an applied software engineering methodology involving requirement analysis, review of existing approaches, system modelling, database design, prototype implementation, and evaluation planning. The reference paper uses a progression from problem motivation and related work to research gap, proposed framework, methodology, implementation, evaluation, limitations, and conclusion; this paper follows the same research pattern while developing original content for the Fake News Detection project. The reference emphasizes separating proposed evaluation from completed experimental