Model Selection: Non-Nested Model Selection

This lecture covers the theory and practice of model selection, focusing on non-nested models. It discusses automatic model selection and building, exploring criteria like AIC, BIC, and Cp statistic. The bias/variance tradeoff and expected prediction error are also explained, along with the concepts of design matrix, true model, and correct/wrong models. Various model selection criteria and methods like cross-validation, AIC, and simulation experiments are presented, providing insights into the tradeoff between model complexity and accuracy.