(A) Observed absolute errors for boosted training set A, which had the best external predictive performance (sPRED = 0.633; sCV = 0.762).
(B) Observed absolute errors for boosted training set B, which had the best internal predictive performance (sCV = 0.681; sPRED = 0.637).
(D) Predicted uncertainties for boosted training set A. (E) Predicted uncertainties for boosted training set B. (F) Predicted uncertainties for the biased training set R. Several conclusions can be drawn by comparing the distribution of errors to each other and to the distribution of activities.
(B) Results for the model constructed from the biased training set R. (C) Results from boosted training set A. (D) Results for boosted training set B. Eq. 7 implies that = |et| for each member t in the training set.
There's been lots of talk of boosting train times between the big northern cities, but for the first time we've got a price list.
The optimal number of components p* for the CoMFA models obtained for the boosted training sets ranged from three to seven.
The distributions of errors for the boosted training sets are much better behaved; indeed, the predicted uncertainties are slightly more conservative than necessary for large errors in prediction (Fig. 5C and 5D).
Hence HQSAR performance followed the trend seen for CoMFA: cross-validation under-estimated the predictive error substantially for the biased subset (i.e., was overly optimistic about the extensibility of the model) and over-estimated the predictive error slightly for the boosted training sets.
Thirdly, the distributions of predictive uncertainty seen for the boosted training sets are in good overall agreement with the observed errors with respect to the regions of descriptor space where the observed error is relatively high or low (Fig. 4D vs 4A and 4E vs 4B).
A total of 113 inhibitors were not selected for any of the boosted training sets, whereas 191 were selected for at least one of them.
