Skip to main content
. Author manuscript; available in PMC: 2014 Jun 15.
Published in final edited form as: Cancer Res. 2013 Apr 10;73(12):3525–3533. doi: 10.1158/0008-5472.CAN-12-4244

Table 2. Summary of cross-validation results.

Upper panel shows the percentage of cases in which the bi-exponential model was superior to the exponential model, when both were fit to the training data sets. Lower panel is the percentage of cases in which the bi-exponential model was superior to the exponential model, as determined by lower value of MSE on the testing data set, for each model determined by fitting to the validation data sets.

Percentage of training data sets for which the bi-exponential model fit was better
Cross-validation According to MSE According to AIC
1 100 % 100 %
2 100 % 98.2 %
3 100 % 97.9 %
4 100 % 75.3 %
5 100 % 100 %
Percentage of validation data sets for which the bi-exponential model fit was better
Cross-validation According to MSE
1 53 %
2 79.2 %
3 80.9 %
4 86.5 %
5 100 %