Related Experiment Video
Updated: Aug 8, 2026
08:20
Superior Auto-Identification of Trypanosome Parasites by Using a Hybrid Deep-Learning Model
Published on: October 27, 2023
Intelligent malware detection on Android smartphones via a hybrid approach using gradient boosting and convolutional
Collins Chimeleze1,2, Norziana Jamil3, Zuhaira Muhammad Zain4
1Institute of Informatics and Computing in Energy, Universiti Tenaga Nasional, Selangor, Malaysia. chimeleze@uniten.edu.my.
Scientific Reports
|August 6, 2026
Summary
This study introduces CNN-GBM, a hybrid model for Android malware detection. It effectively identifies malicious software with improved accuracy and reduced error rates compared to existing deep learning methods.
Area of Science:
- Computer Science
- Cybersecurity
- Machine Learning
Background:
- Malware detection is crucial for system security.
- Existing machine learning methods often require extensive feature engineering, increasing computation time and errors.
- Deep learning models show promise but can be computationally intensive.
Purpose of the Study:
- To propose a novel hybrid model, CNN-GBM, for enhanced Android malware detection.
- To improve the efficiency and accuracy of malware identification systems.
- To overcome limitations of existing deep learning techniques in malware analysis.
Main Methods:
- A hybrid model combining Convolutional Neural Networks (CNN) and Gradient Boosted Machines (GBM) was developed.
- Data was represented using n-grams, opcode sequences, or byte-level features.
- An entropy-based feature selection method was employed before GBM classification.
Main Results:
- The CNN-GBM model demonstrated superior performance over existing deep learning models.
- Key improvements include a 3% increase in accuracy, 5% in recall, and 3.5% in AUC.
- Significant reductions were observed in false positive rate (-52%), false negative rate (-15%), and overall error rate (-35%).
Conclusions:
- The CNN-GBM hybrid model offers a more effective approach to Android malware detection.
- Integrating CNN's feature extraction with GBM's boosting capabilities enhances detection performance.
- This model presents a promising solution for improving cybersecurity against mobile malware.