Corrected error in k fold

ayrna · Jan 20, 2018 · 933a010 · 933a010
1 parent 086dc7b
commit 933a010
Show file tree

Hide file tree

Showing 2 changed files with 52 additions and 2 deletions.
diff --git a/doc/orca-tutorial.md b/doc/orca-tutorial.md
@@ -473,7 +473,57 @@ ans =
     dlmwrite(fullfile(rootDir,sprintf('test-%s.%d',nameDataset,ff-1)),ERAData(teIdx,:),' ');
 end
 ```
-This will generate all the partitions for a `10`fold crossvalidation experimental design. The source code of this example is in [exampleERAKFold.m](../src/code-examples/exampleERAKFold.m). In order to obtain a `30`holdout design, the code will be a bit different:
+This will generate all the partitions for a `10`fold crossvalidation experimental design. The source code of this example is in [exampleERAKFold.m](../src/code-examples/exampleERAKFold.m).
+
+In order to obtain a `30`holdout design, the code will be a bit different. As MATLAB does not included a native way of repeating holdout, we will do it manually:
+```Matlab
+>> % Load data
+ERAData = table2array(readtable('../../exampledata/ERA.csv'));
+>> % Check the first 20 rows
+ERAData(1:20,:)
+
+ans =
+
+     3     2     0    14     1
+     3     3     5     9     1
+     1     3    10     7     1
+     0     5     7     2     1
+     0    10    12     7     1
+     0     5     7     2     1
+     2     1     5     1     1
+     0     5     7     2     1
+     3     2     0    14     1
+    10     7     1     6     1
+     1     2    12     4     1
+     2     1     5     1     1
+     1     3    10     7     1
+     0     5     7     2     1
+     3     2     0    14     1
+     1     9     4     1     1
+     2     1     5     1     1
+     3     2     0    14     1
+     2     1     5     1     1
+     5     7     3    12     1
+
+>> % Extract targets
+targets = ERAData(:,end);
+>> % Generate h holdout partitions
+h=30;
+>> % Prepare filesystem
+nameDataset = 'era';
+rootDir = fullfile('..', '..', 'exampledata', '30-holdout', nameDataset);
+mkdir(rootDir);
+rootDir = fullfile(rootDir,'matlab');
+mkdir(rootDir);
+>> % For each partitions
+for ff = 1:h
+    CVO = cvpartition(targets,'HoldOut',0.25); % 25% of patterns for the test set
+    trIdx = CVO.training(1);
+    teIdx = CVO.test(1);
+    dlmwrite(fullfile(rootDir,sprintf('train-%s.%d',nameDataset,ff-1)),ERAData(trIdx,:),' ');
+    dlmwrite(fullfile(rootDir,sprintf('test-%s.%d',nameDataset,ff-1)),ERAData(teIdx,:),' ');
+end
+```
 
 
 ### Warning about highly imbalanced datasets

diff --git a/src/code-examples/exampleERAKFold.m b/src/code-examples/exampleERAKFold.m
@@ -9,7 +9,7 @@
 
 % Generate k fold partitions
 k=10;
-CVO = cvpartition(targets,'k',k);
+CVO = cvpartition(targets,'KFold',k);
 
 % Prepare filesystem
 nameDataset = 'era';