The message reporting a warning for the ratio of calibrated to validated residual variance appears for a PLS model when the residual variance for Y from the validation is much higher than the calibration. In other words, a 3-factor model explains much less of the Y variance for new samples compared to the training samples. This is an indication that the models performs notably worse when applied to new samples compared to when applied to those used for building the model. Including more than 2 factors might therefore be overfitting the model.

## Thread Reply

## Please briefly explain why you feel this answer should be reported .

Report CancelThe message reporting a warning for the ratio of calibrated to validated residual variance appears for a PLS model when the residual variance for Y from the validation is much higher than the calibration. In other words, a 3-factor model explains much less of the Y variance for new samples compared to the training samples. This is an indication that the models performs notably worse when applied to new samples compared to when applied to those used for building the model. Including more than 2 factors might therefore be overfitting the model.