Recealer: A malware detection method based on machine learning and deep learning models

Authors: Asia Othman Aljahdali *, Elaf Maqadmi, Atouf Ghabashi, Deem Alsuoilme, Bayader Alluhaybi, Edmy Alboqami

Affiliations:

College of Computer Science and Engineering, University of Jeddah, Jeddah, Saudi Arabia

Abstract

Raccoon Stealer malware is difficult to detect as it actively evades traditional methods. This study proposes Recealer, a hybrid detection model that integrates machine learning (ML) and deep learning (DL). The model applies static and dynamic analysis to extract features from sample files, which are first classified using an ML algorithm. Files with uncertain classification results are then transformed into grayscale images and analyzed by a convolutional neural network for improved precision. Experimental results demonstrate that Recealer achieves 94% overall detection accuracy, with the random forest algorithm attaining 97.53% in the ML stage and the DL stage reaching 95%. These findings indicate that Recealer is both efficient and reliable for detecting Raccoon Stealer malware.

Keywords

Malware detection, Machine learning, Deep learning, Hybrid model, Feature extraction

Download

📄 Full PDF

DOI

https://doi.org/10.21833/ijaas.2025.10.005

Citation (APA)

Aljahdali, A. O., Maqadmi, E., Ghabashi, A., Alsuoilme, D., Alluhaybi, B., & Alboqami, E. (2025). Recealer: A malware detection method based on machine learning and deep learning models. International Journal of Advanced and Applied Sciences, 12(10), 36–44. https://doi.org/10.21833/ijaas.2025.10.005