Code and Results for Chapter 3

• Code that duplicates the numerical results from the text

  • Duplicates the numerical results in Tables 3.1 and 3.2 from the book
    • duplicate_table_3_1_N_2.R
  • Duplicate figure 3.5 from the book (the all possible subsets model)
    • duplicate_figure_3_5.R
  • Duplicate figure 3.7 and table 3.3: (linear methods with the prostate data set)
    • All Subsets:
      • dup_OSE_all_subsets.R (duplicates all subsets cross validation)
      • cv_all_subsets.R (cross validation of all subset models)
    • Ridge Regression:
      • dup_OSE_ridge_regression.R (duplicates ridge regression)
      • cv_Ridge.R (cross validation of ridge regression results)
      • opt_lambda_ridge.R (Returns the "optimal" set of lambda for cross validation)
      • df_ridge.R (implements the degrees of freedom for ridge regression)
    • Lasso:
      • dup_OSE_lasso.R (duplicates lasso regression)
        
    • Principal Component Regression and Partial Least Squares:
      • dup_OSE_PCR_N_PLSR.R (duplicates PCR and PLS regressions)
        
      • cv_PCR_N_PLSR.R (cross validation for PCR and PLS)
        
      • pcr_wwx.R (principal component regression from scratch)
        
      • cv_pcr_wwx.R (cross validation using pcr_wwx)
        
      • pcr_wwx_run.R (runs the prostate data set using pcr_wwx)
        
    • Incremenal Forward Stagewise Regression (IFSR):
      • IFSR.R (duplicates IFSR)
        
      • IFSR_run.R (runs IFSR and duplicates figure 3.9)
        
    • Auxiliary routines:
      • load_prostate_data.R (load the prostate data and optionally applies simple data scaling)
        
      • one_standard_error_rule.R (implements the one standard error rule)
        
  • Duplicates figure 3.8 from the book (the ridge regression coefficients)
    
    • duplicate_figure_3_8.R

• Code for Exercises:

  • Ex3.2.R (code to solve Exercise 3.2)
    
  • prob_3_16.R (code used in Exercise 3.16)
    
  • Problem 3.17 (linear methods applied to the SPAM data set)
    
    • load_spam_data.R (load and optionally scales the SPAM data set)
      
    • spam_OLS.R (ordinary least squares regression)
      
    • spam_ridge.R (applies ridge regression)
      
    • spam_lasso.R (applies the lasso penalty term to beta)
      
    • spam_PCR_N_PLS.R (principal component regression and partial least squares regression)