The objective of this course is to equip students with the knowledge and tools to understand, implement, and critically assess machine learning methods in the context of finance and insurance. The course emphasizes the integration of theoretical foundations with empirical modeling and the application of machine learning techniques to real-world problems such as forecasting, risk modeling, and portfolio optimization.

Objectives of the course are set in a way that students will:
- Understand core concepts of machine learning (e.g., supervised vs. unsupervised learning, overfitting, model validation);
- Be able to select and justify appropriate models based on data characteristics and analytical objectives;
- Apply machine learning tools to solve practical problems in finance (e.g., return prediction, credit scoring, volatility modeling);
- Develop practical skills in working with real datasets, implementing models in programming environments (e.g., Python, R), and interpreting and communicating model results.

Course will enable students to develop the following competences:
- Explain key machine learning methods and their advantages and limitations in financial applications;
- Apply linear and non-linear models for regression, classification, and clustering tasks;
- Perform model validation and critically assess model performance using appropriate metrics (e.g., AUC, RMSE, out-of-sample error);
- Integrate methods such as regularization, tree-based algorithms, principal component analysis, and clustering into quantitative financial analysis;
- Design and analyze models for return forecasting, volatility estimation, credit risk scoring, and portfolio optimization under uncertainty;
- Understand the importance of model interpretability and robustness in financial decision-making;
- Synthesize knowledge of statistics, finance, and programming to address complex problems in actuarial science and finance.

- Introduction to machine learning: supervised vs. unsupervised learning, predictive vs. causal inference, challenges in finance.
- Linear regression: model estimation, interpretation, and diagnostics.
- Train-test split, cross-validation, and the bias-variance trade-off.
- Classification methods: logistic regression, linear discriminant analysis, and evaluation metrics (accuracy, precision, recall, AUC).
- Model selection: best subset selection, AIC/BIC, cross-validation.
- Regularization methods: ridge and lasso regression (with applications in portfolio construction).
- Resampling methods: cross-validation, bootstrap, and walk-forward analysis for time series.
- Tree-based methods: decision trees, random forests, and boosting.
- Clustering techniques: K-means, hierarchical clustering, Principal Component Analysis (factor models, risk decomposition).
- Support Vector Machines (basic concepts and classification applications).
- Neural networks (basics only): feedforward architecture, overfitting, and regularization.
- Model interpretability and robustness: feature importance, SHAP values, and explainability in financial contexts.
- Applications in finance: return prediction, volatility forecasting, credit scoring, and portfolio optimization.