Identifying Rumor Sources in Social Networks Using Hashing Vectorizer for Text Representation

In today's digitally connected world, the rapid dissemination of rumors and misinformation across social networks presents significant challenges to public trust and information integrity. This study introduces a novel approach for Rumor Source Identification (RSI) using text analytics and network analysis techniques. We employ the Hashing Vectorizer for efficient and scalable text representation, enabling us to process large volumes of social media data. For classification, we utilize the Gaussian Naive Bayes algorithm to determine whether a message constitutes a rumor or not. Our methodology focuses on tracking the diffusion patterns of information and identifying the original sources responsible for initiating rumor cascades. The frontend of our system is developed using HTML, CSS, and JavaScript, while the backend is implemented with Flask, enabling a user-friendly and interactive platform for visualization and analysis. This research contributes to the development of effective tools for early rumor detection and source tracing, ultimately aiming to mitigate the spread of misinformation and enhance the reliability of online information.