4/21/2017
Exam 2
- Coverage: Muliple Regression Models, Chapters 7-11
- Samples are provided in P:\data\math\hartlaub\regression\sample exams
Chapter 7
- Extra Sums of Squares
- Using extra sums of squares to test hypotheses
- Coefficients of partial determination
- Standardized Multiple Regression
- Full versus Reduced Model F tests
- Multicollinearity
Chapter 8
- Polynomial regression
- Models with interaction
- Contrained regression
- Regression with qualitative predictors, using indicator variables
Chapter 9
- Best subsets regression
- Automatic search procedures - forward, backward, and stepwise regression models
- Criteria for evaluating different models (R-squared, adjusted R-squared, SSE, MSE, Mallow's Cp, AIC, SBC, and PRESSp
- Model validation
Chapter 10
- Added variable plots
- Rules of thumb for identifying outliers based on studentized residuals, deleted residuals, and studentized deleted residuals
- Using the hat matrix and leverage to indentify outlying X observations
- Rules of thumb for identifying influential cases with DFFITS, Cook's Distance, and DFBetas
- Multicollinearity diagnostics - rules of thumb for VIFs
Chapter 11
- Weighted least squares
- Ridge regression
- Least absolute regression or (L1 norm regression)
- Least median squares
- Iteratively reweighted least squares
- LOESS
- Bootstrapping
- Other methods (spline regression, etc.)