Machine Failure Prediction Using Random Forest

📌 Overview

This project focuses on predicting machine failures using sensor data. The goal is to analyze various factors such as temperature, vibration, pressure, and humidity to determine if a machine is likely to fail.

📂 Dataset

• Features:
  • Temperature(°C)
  • Vibration(Hz)
  • Pressure(bar)
  • Humidity(%)
  • Timestamp (Categorical)
  • Machine_ID (Categorical)
  • Failure_Status (Target Variable: 0 - No Failure, 1 - Failure)

🛠️ Data Preprocessing & Feature Engineering

• Handling Missing Data:
  • Dropped columns with more than 50% missing values
  • Filled missing numerical values with the median
  • Filled missing categorical values with the mode
• Feature Encoding:
  • Converted categorical columns using Label Encoding & One-Hot Encoding
• Scaling & Normalization:
  • Applied StandardScaler to numerical features

📊 Model Training & Tuning

• Baseline Model: Random Forest Classifier
• Hyperparameter Tuning: Performed GridSearchCV with 5-fold cross-validation
• Best Parameters Found:

  max_depth: 10
  min_samples_leaf: 1
  min_samples_split: 5
  n_estimators: 100
  

• Final Model Performance:

  Metric	Value
  Accuracy	97%
  Precision	99% (Class 0), 82% (Class 1)
  Recall	98% (Class 0), 90% (Class 1)
  F1-Score	98% (Class 0), 86% (Class 1)

📈 Results & Insights

• The model performs well, achieving 97% accuracy.
• Precision and recall are high, meaning the model can effectively predict failures.
• Feature importance analysis (SHAP) showed Vibration and Temperature as key factors.
• Confusion matrix & ROC curve analysis confirmed the model's reliability.

🚀 Deployment (Optional)

This model can be deployed as:

1. Streamlit Web App for interactive predictions
2. Flask/FastAPI API for integration with other systems
3. Dockerized & Deployed on Cloud (AWS/GCP/Azure)

📌 Future Improvements

🔹 Try deep learning models like LSTMs for time-series predictions.
🔹 Optimize feature selection for even better performance.
🔹 Deploy the model in real-time industrial settings.

📜 References

• Sensor failure prediction research papers
• Random Forest algorithm documentation
• SHAP for interpretability