<![CDATA[Given the rise of increasingly sophisticated malware, it is important to develop accurate and fast detection systems to minimize the losses incurred. This study adopts a hybrid approach by integrating the Random Forest algorithm for malware classification and the TOPSIS method for prioritizing malware handling based on risk factors and business impact. The dataset used is from CIC-MalMem-2022 and processed using statistical features such as entropy, number of API calls, and file size. The Random Forest model was optimized using Grid Search, with the best results at parameters n_estimators = 100 and max_depth = 10, reaching a detection accuracy of 95.87% after hyperparameter tuning. Subsequently, the decision-making process was conducted using the TOPSIS method to rank malware based on predefined criteria weights. Evaluation results show that this system achieves a success rate with a prioritization accuracy of 0.84 and detection and response times under 30 minutes, thereby supporting more effective cybersecurity needs. Thus, this approach has proven capable of significantly improving malware detection accuracy and accelerating the mitigation process.]]>
Copyrights © 2025
