Tweak default values of maxit and epsilon

R/abclass.R

@@ -125,9 +125,9 @@ abclass <- function(x, y,
 ##' @param boost_umin A negative number for adjusting the boosting loss for the
 ##'     internal majorization procedure.
 ##' @param maxit A positive integer specifying the maximum number of iteration.
-##'     The default value is \code{10^5}.
+##'     The default value is \code{10^4}.
 ##' @param epsilon A positive number specifying the relative tolerance that
-##'     determines convergence.  The default value is \code{1e-6}.
+##'     determines convergence.  The default value is \code{1e-7}.
 ##' @param standardize A logical value indicating if each column of the design
 ##'     matrix should be standardized internally to have mean zero and standard
 ##'     deviation equal to the sample size.  The default value is \code{TRUE}.
@@ -154,8 +154,8 @@ abclass.control <- function(lambda = NULL,
                             lum_a = 1.0,
                             lum_c = 0.0,
                             boost_umin = - 5.0,
-                            maxit = 1e5L,
-                            epsilon = 1e-6,
+                            maxit = 1e4L,
+                            epsilon = 1e-7,
                             standardize = TRUE,
                             varying_active_set = TRUE,
                             verbose = 0L,

R/abrank.R

@@ -51,8 +51,8 @@ abrank.control <- function(lambda = NULL,
                            lum_a = 1.0,
                            lum_c = 1.0,
                            boost_umin = - 5.0,
-                           maxit = 1e5L,
-                           epsilon = 1e-4,
+                           maxit = 1e4L,
+                           epsilon = 1e-7,
                            standardize = TRUE,
                            varying_active_set = TRUE,
                            verbose = 0L,

inst/include/abclass/Control.h

@@ -48,8 +48,8 @@ namespace abclass
         //   elastic-net
         double alpha_ { 1.0 };
         //   group {scad,mcp}
-        double kappa_ratio_ { 0.99 }; // parameter to set gamma
-        double gamma_ { - 1.0 };      // gamma for group non-convex penalty
+        double kappa_ratio_ { 0.9 }; // parameter to set gamma
+        double gamma_ { - 1.0 };     // gamma for group non-convex penalty
 
         // tuning
         //   cross-validation
@@ -61,8 +61,8 @@ namespace abclass
         unsigned int et_nstages_ { 0 };
 
         // optimization
-        unsigned int max_iter_ { 100000 }; // maximum number of iterations
-        double epsilon_ { 1e-6 };          // tolerance to check convergence
+        unsigned int max_iter_ { 10000 };  // maximum number of iterations
+        double epsilon_ { 1e-7 };          // tolerance to check convergence
         bool varying_active_set_ { true }; // if active set should be adaptive
         bool standardize_ { true };        // is x_ standardized (column-wise)
         unsigned int verbose_ { 0 };