Detection of Malicious Cyber Fraud using Machine Learning Techniques

Abstract

As the technology and internet have come to their dawn, the rate of cyber-crimes has also increased. This increases the risk of information insecurity and the spread of crimes such as spam, farming and phishing, financial fraud, etc. Particularly, the attackers/hackers spread malicious uniform resource locators (URLs) to exploit vulnerabilities of the system and gain the personal information of the users. Thus, a study on malicious URL detection is necessary to prevent such attacks. Several studies exist which show numerous ways to determine malicious URLs based on machine learning (ML) and deep learning (DL), but there are some problems, for example, malicious features cannot be extracted efficiently. In this research, a model is proposed to ascertain malicious URLs, which is formulated on random forest, support vector machine (SVM), deep neural network (DNN), convolutional neural network (CNN). The several datasets are considered containing malicious and benign URLs to train the model to detect URL behaviour and attributes. The empirical results show that the suggested method can detect malicious URLs efficiently, based on URL behaviour and attributes. Thus, the solution may be advised as an efficient and reliable solution for the problem of malicious URL detection.