External validation
Posted: Mon Sep 09, 2019 3:10 pm
Hello Everyone,
I'd like to know how to conduct an external validation of a model ( on a different dataset from those used for model development). I'm aware of the PLSpredict algorithm, but I need to carry out an external validation instead of a "holdout-sample-based procedure " or cross-validation procedure. Which metric should I look for to compare the two models? Which criteria define a "good" predictive power in this case?
Thank you very much,
Alessandra
I'd like to know how to conduct an external validation of a model ( on a different dataset from those used for model development). I'm aware of the PLSpredict algorithm, but I need to carry out an external validation instead of a "holdout-sample-based procedure " or cross-validation procedure. Which metric should I look for to compare the two models? Which criteria define a "good" predictive power in this case?
Thank you very much,
Alessandra