Skip to main content
. 2007 Feb 16;2:133–162.

Figure 7.

Figure 7.

Pictorial simplified example of nested 3-fold cross-validation. The data are split into 3 mutually exclusive sets of samples: P1, P2 and P3. The performance is estimated in the outer loop by training on all sample sets but one, using the remaining one for testing. The average performance over testing sets is reported. The inner loop is used to determine the optimal value of the classifier’s parameter C that takes values “1” and “2” (in a cross-validated fashion). This value of parameter C is used for training in the outer loop. A detail algorithmic description is provided in [Statnikov, 2005b].