Goal of this Project
Predict Ransomware & Malware based on file properties extracted from a tool. It's a classification problem (Supervised Machine Learning). The data was imbalanced and must be transformed using (Synthetic Samples: SMOTE-Tomek).
Highlights
- LazyPredict for AutoML Official Documentation
- LIME for Local Explainations
- Weight of Evidence (Feature Selection Technique on Feature Separation Power) Read More
 |
|---|
| LIME Explainability for Local Interpretation |
Model Performance on Test Dataset
 |
|---|
| Confusion Matrix |
Metrics Test Data Results:
- Model Used: Random Forest
- Accuracy: 0.9933
- Precision: 0.9847
- Recall: 0.9931
- F1 Score: 0.9889
- MCC: 0.9841
- False Positive Rate: 0.0067
- AUC Score: 0.9994
Install Libraries using requirements.txt
pip install -r /path/to/requirements.txt