Statistical modeling foundations for machine learning (MSc level). The course is organized in short lecture segments followed by hands-on practical sessions.

Includes a visual presentation of artificial neural networks created in Manim in collaboration with Lucas Robinet (doing remarkable work).

Probability & Estimation Basics

• Random variables, probability density, and core probability notions
• Empirical estimation of model parameters
• Maximum likelihood as a unifying viewpoint

Linear Regression as a Statistical Model

• Simple (1D) linear regression: model formulation and error analysis
• Multiple linear regression and statistical inference
• Outlier detection and diagnostics

Model Selection & Generalization

• Overfitting and the need for validation
• Cross-validation workflows
• Regularization: Ridge and LASSO
• Bias–variance trade-off and curse of dimensionality (intuition + practice)

Statistical Testing & Structured Linear Models

• ANOVA: building and interpreting statistical tests
• Mixed (linear) models: parameter estimation and factor significance testing

Openings Toward Classic ML Linear Methods

• Logistic regression: mathematical construction + interpretability questions
• Partial Least Squares (PLS): dimensionality reduction for regression