ADVANCED DETECTION OF FRAUDULENT AND MALICIOUS URLS BASED ON MACHINE LEARNING

With the escalating risk of cyber fraud driven by sophisticated techniques like social engineering and phishing,the detection of malicious Uniform Resource Locators (URLs) has become imperative. This paper introduces anovel approach to detecting cyber fraud, specifically focusing on malicious URLs and text by leveragingmachine learning techniques and big data analytics. Traditional methods such as signature-based detectionsuffer limitations in identifying new threats, prompting the need for more advanced methodologies. Theproposed system utilizes a combination of machine learning algorithms, namely Support Vector Machine(SVM), Naive Bayes, Logistic Regression, Decision Tree Classifier, SGD Classifier, to classify URLs and textmessages based on their behaviors and attributes. Unlike existing systems, this approach incorporates both staticand dynamic characteristics of URLs, leading to a more comprehensive detection mechanism. The novelty liesin the introduction of new URL features specifically tailored for malicious URL and text detection. Advantagesof the proposed system include its ability to harness the potential of these new features, enhancing the accuracyand efficiency of malicious URL and text detection. While These are showcased in this research, the frameworkremains flexible, allowing for the incorporation of other algorithms. Experimental results prove the efficiency ofthe proposed method in significantly improving malicious URL detection and Text detection making it apromising solution in combating cyber fraud.