To achieve more stable cross-validation results, any possible effect of a particular random split needs to be avoided.
SRCV leads to more stable decisions for both validation and selection which are not biased by underlying biological phenomena.
PLSR and PCR proved to be more stable in a nested cross-validation protocol.
Note that the results of multiple imputation distributions are clear: the multiple imputation not only improves upon single imputation, by providing better validation statistics, but the validation statistics are much more stable than those from a single imputation [ 18].
Multiple outperforms single imputation by providing better and much more stable estimates, according to the computed validation statistics.
Djabalah also presented preliminary data showing that iPS cells provide a more stable platform for drug discovery in terms of target identification and validation for both inhibitors and drivers of differentiation.
By doing repeated cross-validation, it is possible to reduce variability and provide more stable estimates.
In summary, a cross-validation in a larger and more robust corpus could provide more stable results.
The split-sample validation approach suggested here tries to mitigate the last issue, and while not strictly a validation, it may be used as a robustness check as true interactions should be more stable in subsets of data than false ones.
