added BOOTDATA as third output argument to bootclust

inst/bootclust.m

@@ -92,17 +92,22 @@
 %     so that bootclust results are reproducible.
 %
 %     'bootclust (DATA, NBOOT, BOOTFUN, ALPHA, ..., LOO, SEED, NPROC)' also
-%     sets the number of parallel processes to use for function evaluations
-%     during bootstrap and jackknife computations on multicore machines. This
-%     feature requires the Parallel package (in Octave), or the Parallel
-%     Computing Toolbox (in Matlab).
+%     sets the number of parallel processes to use for jackknife computations
+%     and non-vectorized function evaluations during bootstrap and on multicore
+%     machines. This feature requires the Parallel package (in Octave), or the
+%     Parallel Computing Toolbox (in Matlab). This option is ignored during
+%     bootstrap function evaluations when BOOTFUN is vectorized.
 %
 %     'STATS = bootclust (...)' returns a structure with the following fields
 %     (defined above): original, bias, std_error, CI_lower, CI_upper.
 %
 %     '[STATS, BOOTSTAT] = bootclust (...)' returns BOOTSTAT, a vector or matrix
 %     of bootstrap statistics calculated over the bootstrap resamples.
 %
+%     '[STATS, BOOTSTAT, BOOTDATA] = bootclust (...)' returns BOOTDATA, a 1-by-
+%     NBOOT cell array of datasets generated by cluster or block bootstrap
+%     resampling.
+%
 %  REQUIREMENTS:
 %    The function file boot.m (or better boot.mex) and bootcdf, which are
 %    distributed with the statistics-resampling package.
@@ -140,8 +145,8 @@
 %  You should have received a copy of the GNU General Public License
 %  along with this program.  If not, see http://www.gnu.org/licenses/
 
-function [stats, bootstat] = bootclust (x, nboot, bootfun, alpha, clustid, ...
-                                        loo, seed, ncpus)
+function [stats, bootstat, X] = bootclust (x, nboot, bootfun, alpha, ...
+                                        clustid, loo, seed, ncpus)
 
   % Store local functions in a stucture for parallel processes
   localfunc = struct ('col2args', @col2args, ...
@@ -159,7 +164,7 @@
   if (nargin > 8)
     error ('bootclust: Too many input arguments')
   end
-  if (nargout > 2)
+  if (nargout > 3)
     error ('bootclust: Too many output arguments')
   end
 
@@ -407,7 +412,7 @@
     M = cell2mat (cellfun (@(i) repmat (x(:, i), 1, 2), ... 
                  num2cell (1 : nvar), 'UniformOutput', false));
   else 
-    M = repmat (x, 1, 2); 
+    M = repmat (x, 1, 2);
   end
   if (any (szx > 1))
     VECTORIZED = false;
@@ -442,30 +447,27 @@
 
   % Perform resampling of clusters
   bootsam = boot (nx, nboot, loo);
-  X = arrayfun (@(i) x(bootsam(i, :)), 1 : nx, 'UniformOutput', false);
-  X = [X{:}]';
+  X = arrayfun (@(b) cell2mat (x(bootsam(:, b))), 1 : nboot, ...
+                     'UniformOutput', false);
 
   % Perform the function evaluations
-  if VECTORIZED
-    X = reshape (vertcat (X{:}), n, nboot * nvar);
-    bootstat = bootfun (X);
+  if ( (VECTORIZED) && (nvar == 1) )
+    bootstat = bootfun (cell2mat (X));
   else
     if (ncpus > 1)
       % Evaluate bootfun on each bootstrap resample in PARALLEL
-      X = arrayfun (@(i) vertcat (X(:,i)), 1 : nboot, 'UniformOutput', false);
       if (ISOCTAVE)
         % OCTAVE
-        bootstat = parcellfun (ncpus, @(X) bootfun (cell2mat (X)), X, ...
+        bootstat = parcellfun (ncpus, @(x) bootfun (x), X, ...
                                'UniformOutput', false);
-
       else
         % MATLAB
         bootstat = cell (1, nboot);
-        parfor b = 1 : nboot; bootstat{b} = bootfun (cell2mat (X{:, b})); end
+        parfor b = 1 : nboot; bootstat{b} = bootfun (X{:, b}); end
       end
     else
       % Evaluate bootfun on each bootstrap resample in SERIAL
-      bootstat = cell2mat (arrayfun (@(b) bootfun (vertcat (X{:,b})), ...
+      bootstat = cell2mat (arrayfun (@(b) bootfun (X{:, b}), ...
                                           1 : nboot, 'UniformOutput', false));
     end
   end

inst/bootknife.m

@@ -939,8 +939,8 @@ function print_output (stats, nboot, alpha, l, m, bootfun_str, strata)
 
 function retval = col2args (func, x, szx)
 
-  % Usage: retval = col2args (func, x, nvar)
-  % col2args evaluates func on the columns of x. When nvar > 1, each of the
+  % Usage: retval = col2args (func, x, szx)
+  % col2args evaluates func on the columns of x. When sum(szx) > 1, each of the
   % blocks of x are passed to func as a separate arguments. 
 
   % Extract columns of the matrix into a cell array