MULTI MODEL CYBER THREAT DETECTION SYSTEM USING MACHINE LEARNING

Abstract

In the digital age, ensuring internet security has become critical. This research describes an
innovative technique to mitigating cyber dangers using a multi-modal detection system integrated in a webpage.
The system has three separate functionalities: email spam detection, phishing mail detection, and cyberbullying
detection. Each operation is encased under a distinct button on the webpage interface, allowing for easy
interaction. When a certain detection job is chosen, the matching model is activated, using advanced machine
learning algorithms trained on relevant datasets. or cyberbullying detection, users submit text material into the
designated field, and the model swiftly evaluates whether the text contains elements indicative of cyberbullying
behavior. Similarly, the email spam and phishing mail detection functions scan incoming emails for distinctive
patterns associated with spam or phishing attempts, protecting users from dangerous information. This multimodal
method provides a comprehensive defense mechanism against a wide range of cyber threats. This study
gives a thorough evaluation of the most recent state-of-the-art strategies for detecting cyberbullying on social
media platforms using machine learning. We explore several machine learning models, feature extraction
approaches, and datasets used for cyberbullying detection, which improves online security and promotes a better
digital environment for consumers globally.