Trade-off between Interpretability and Predictive Accuracy in Regression Tasks

In modern data analysis, it is often necessary to choose between models that are easy to interpret, such as linear and logistic regression, and models that offer higher predictive accuracy, such as decision trees. In many application areas, including medicine, transportation, and social sciences, interpretability is an important requirement. The aim of this thesis is to compare these types of models using real data. The thesis focuses on interpreting model outputs, evaluating predictive performance using validation methods, and discussing the practical consequences of model choice. Special attention is paid to model interpretability and their usability in applied problems.