PhishGuard: Machine Learning Model for Real-Time URL Detection

The identification of phishing URLs has become essential to cybersecurity due to the increasing complexity of cyber threats. Phishing attacks are a serious risk to both persons and companies because they contain misleading web links that are meant to fool users into disclosing important information. This paper presents it. We employed the Multinomial Naïve Bayes and Logistic Regression algorithms in our study; comparing among the two algorithms we choose Logistic Regression algorithm because it is best suited for text classification and its accuracy in detecting Phishing URL is better compared to the multinomial Naïve Bayes algorithm. The project is integrated with a user interface page where a user can enter the URL or upload an image that containing the URL text to predict whether that URL is Phish URLs or legitimate URL. We extract the text from the image using methods that process images, and then we utilize Regex to retrieve the URL. Additionally, we have created web applications using Flask, routing URL patterns and mapping them to Python functions or scripts.