# Predictive Maintenance – Engine Condition Model

This repository contains the **best-performing model** trained on engine sensor data
to predict the target: `Engine_Condition`.

## ✅ Best Model
- Algorithm: **GradientBoosting**
- Test F1-score: **0.7601**
- Test Accuracy: **0.6657**
- Best Parameters: `{"clf__learning_rate": 0.1, "clf__max_depth": 3, "clf__n_estimators": 100}`

## 📊 Features Used
- Engine_RPM
- Lub_Oil_Pressure
- Fuel_Pressure
- Coolant_Pressure
- Lub_Oil_Temperature
- Coolant_Temperature

## 🔧 Training Setup
- Models tried: DecisionTree, RandomForest, GradientBoosting
- Cross-validation: 3-fold
- Scoring metric: F1
- Preprocessing: Median imputation + StandardScaler