Part quality predictor

The brief

Inspection at the end of the line finds bad parts too late. Predict from process measurements whether each part will pass, so problems are caught earlier, and show which measurements matter most. Good looks like: careful handling of hundreds of features and missing values, metrics for rare failures, and the key signals named.

What you will build

1. Clean the features and handle missing values
2. Reduce features with selection or PCA
3. Train classification models and compare them
4. Evaluate with recall and precision on failures
5. Name the process signals that matter most

Free, open data to build it with

• SECOM semiconductor manufacturing ↗ UCI Machine Learning Repository
• Bosch production line performance ↗ Kaggle

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.