Skip to main content

Supervised learning

A complete study of supervised learning, from linear regression to ensemble methods.

Course Duration: 7h​

What You'll Learn​

  • Choose between regression and classification depending on the question
  • Train and tune linear then logistic regression
  • Understand why a single tree overfits and how an ensemble fixes it
  • Set up reliable cross-validation with no data leakage
  • Choose the right metric when classes are imbalanced

Prerequisites​

  • Python and pandas
  • Notions of descriptive statistics

Course Modules​

  1. Framing a supervised problem: target, features, dataset
  2. Linear regression and Ridge/Lasso regularization
  3. Logistic regression and the decision boundary
  4. k-nearest neighbors and support vector machines
  5. Decision trees: reading and limits
  6. Random forests and bootstrap aggregation
  7. Gradient boosting: XGBoost and LightGBM
  8. Cross-validation, time splits and data leakage
  9. Metrics: accuracy, precision, recall, F1, ROC AUC
  10. Hyperparameter tuning and an end-to-end project

Assessment and certificate​

The course ends with a 40-question exam covering every module. On success, a certificate of completion is issued; its number is verifiable on the platform.

Free courses, by contrast, end with a 5-question quiz and a preview of the certificate, without certification.