Engine Fault Detection Model
Model Overview
This model predicts engine condition (Active / Faulty) using sensor data.
Algorithm
XGBoost Classifier
Training Details
- Objective: Maximize Recall
- Class Imbalance: Handled via class weights / scale_pos_weight
- Outlier Treatment: Coolant temperature capped using training data thresholds
Dataset
- Source: ShanRaja/Data
- Splits: Train / Validation / Test
Metrics (Validation)
- Recall: 1.0
- Precision: 0.63
- F1-score: 0.77
Intended Use
Early fault detection to minimize false negatives.
