A simple Supervised Learning Lab.
I asked GitHub Copilot to quickly scaffold an example using these Supervised Learning (labeled data) techniques:
- XGBoost
- Random Forest
- Gradient Boosting
- Linear Regression
- Logistic Regression
- Ensemble Learning
- Time-Series Forecasting
Start with initial data (randomized, CSV format) then demonstrate the following scenarios:
- Data Leakage
- Underfitting
- Overfitting
- Data Drift
- Concept Drift
- Skew
The above frequently show up in the GCP PMLE but there aren't that many hands-on labs for these in skills.google.
GitHub Copilot diligently implemented this (very literally) and created a Docker example in sequential stages: generate random Data, train the specified Models, then demonstrate scenarios like Data Drift.
docker compose upNote that Data, Models, Metrics will be recreated within the Container. I've included generated examples in the source for study purposes.