One of the world's biggest issues nowadays is cyber-attacks. Every day, they wreak serious economic harm to both persons and nations. Cybercrime is also on the rise along with cyber-attacks. Understanding attack tactics and identifying cybercrime perpetrators are crucial in the fight against crime and criminals. Cyber-attack detection and prevention are challenging undertakings. Yet, academics have lately developed security models and made predictions using artificial intelligence techniques to solve these issues. There are many ways for predicting crimes that can be found in the literature. On the other hand, they struggle to foresee the strategies used in cybercrime and cyber-attacks. By using actual data to pinpoint an assault and its perpetrator, this issue can be solved. The information includes the kind of crime, the perpetrator's gender, the damage, and the attack techniques. Applications made by people who were subject to cyber-attacks can provide the forensic units with data. In this study, we use machine learning to examine two alternative models of cybercrimes and estimate the impact of defined variables on the identification of the attack vector and the perpetrator. In our methodology, we employed eight machine learning techniques and found that their accuracy rates were comparable. With an accuracy percentage of 95.02%, the Support Vector Machine Linear was proven to be the most effective cyber-attack technique. With a high degree of accuracy, the first model allowed us to forecast the types of attacks that the victims were most likely to experience.