Advanced Regression and Modern Statistical Modeling
A systematic examination of how classical OLS assumptions fail in real-world data and what methods replace them. Organized around the failure-modes-of-OLS framework, each module identifies a specific violated assumption and introduces the appropriate advanced estimator or technique. Topics include robust inference, mixed-effects models, generalized least squares, non-linear smoothing via splines and GAMs, quantile regression, instrumental variables with causal graphs, and regularization. Real datasets (Wage, sleepstudy, Ames Housing, Grunfeld, Card, NHANES) are used throughout.
- Level: advanced
- Duration: 14-16 weeks
- 10 sections
- Price: $19.99
Topics: regression analysis, OLS assumptions, GLS, FGLS, mixed effects models, multilevel models, GAMs, splines, quantile regression, instrumental variables, 2SLS, causal inference, DAGs, LASSO, Ridge regression, panel data, robust standard errors, econometrics, R, Python
What's inside
- Advanced Regression and Modern Statistical Modeling
- Chapter 1: OLS Failure Modes, Diagnostics & Inferential Framework
- Chapter 2: GLS, Robust Inference & Panel Data Connections
- Chapter 3: Mixed-Effects (Multilevel) Models & GEE
- Chapter 4: Non-Linearity, Basis Expansions & Smoothing
- Chapter 5: Quantile Regression & Robust Statistics
- Chapter 6: Endogeneity, Causal Graphs & Instrumental Variables
- Chapter 7: Regularization, Geometry & High-Dimensional Inference
- Chapter 8: Model Selection, Validation & Synthesis
- Chapter 9: Conclusion — Where to Go From Here