Nested Model Selection

This lecture covers the concept of nested model selection in linear models, focusing on comparing different models through residual sums of squares and ANOVA. The instructor explains the geometry of linear models, distributions of sums of squares, and the analysis of variance. Through examples with cement data and MPG vs horsepower, the lecture demonstrates how to assess the significance of adding extra variables to a model. The effect of orthogonality in model terms is also discussed, highlighting the uniqueness of sum of squares reductions. Various statistical tests and interpretations are presented to guide model selection decisions.