I am reading a paper which, for different datasets, repeated 5-fold cross validation 10 times on each one of them. For each dataset they calculated the mean of the metric and its standard deviation.

In the end, they provide an average value of the means over the different (unrelated) datasets and an 'average' value of the standard deviations. How is such an 'average' standard deviation value to be computed over the different datasets from the initial CV-folds and how can it be interpreted?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.