comparison core/tools/machine_learning/cvpartition_trunctrain.m @ 0:e9a9cd732c1e tip

first hg version after svn

+% ---
+% class cvpartition_trunctrain
+% NOTE: this is a fake cvpartition double for 
+% using cvpartitions in truncated-training size experiments
+% ---
+classdef cvpartition_trunctrain
+  
+properties (Hidden)
+
+    mtest;
+    mtraining;
+end
+properties
+    N;
+    NumTestSets;
+    TrainSize;
+    TestSize;
+end
+    
+        
+methods
+    
+% ---
+% constuctor: directly calculates the truncated testset
+% ---
+function P = cvpartition_trunctrain(Pin, perctrain)
+  
+    P.N = Pin.N;
+    P.NumTestSets = Pin.NumTestSets;
+    
+    for i = 1:Pin.NumTestSets
+
+        % copy testing data
+        P.TestSize(i) = Pin.TestSize(i);
+        P.mtest{i} = Pin.test(i);
+
+        % calculate new training size
+        P.TrainSize(i) = ceil(perctrain * Pin.TrainSize(i));
+
+        % get actual training indices
+        idx = find(Pin.training(i));
+    
+        % ---
+        % TODO: save the permutation in a global variable,
+        % tomake the same smaller set available
+        % for all further experiments.
+        % moreover, it would be great if the smaller training sets
+        % are subsets of the bigger ones
+        % ---
+        tokeep = randperm(numel(idx));
+        tokeep = tokeep(1:P.TrainSize(i));
+        
+        % get indices to keep
+        idx = idx(tokeep);
+
+        % build truncated training set
+        P.mtraining{i} = false(P.N, 1);
+        P.mtraining{i}(idx) = true;
+    end
+end   
+
+function out = test(P, i)
+
+    out = P.mtest{i};
+end
+
+function out = training(P, i)
+
+    out = P.mtraining{i};
+end
+
+end  
+end