Malware Detection Using Multi-Layer Perceptron Optimized by GridSearch

The identification of malware continues to be a significant challenge in the ever-evolving field of cybersecurity(CS). Traditional signature-based techniques often fall short of recognizing novel and complex threats. This study presents an innovative approach to malware detection by leveraging the capabilities of a Multi-layer Perceptron (MLP) classifier, optimized through GridSearchCV. Unlike conventional techniques, our method integrates machine learning with evolutionary computation to enhance detection accuracy. The MLP classifier, with its deep learning architecture, is adept at capturing intricate patterns within the data, which are crucial for distinguishing between benign and malicious software. However, fine-tuning its hyperparameters is pivotal for achieving optimal performance. The GridSearch-optimized MLP classifier undergoes extensive training, with its hyperparameters continuously refined to maximize performance. Our experimental results indicate a significant improvement in classification performance, showcasing the potential of this combined approach. This research not only underscores the efficacy of integrating MLP classifiers with GridSearchCV for malware detection but also opens avenues for future exploration in optimization techniques. This study contributes a novel methodology to the CS domain, promising enhanced protection against evolving malware threats.