With insufficient predictors, prediction models in this study can only have fair performance.
First, the number of chains in the evaluation set was somewhat larger, since for some interface predictors, predictions could be obtained after some re-formatting of the structure, which was done after development.
For comparison, we downloaded both the other predictors' predictions and our submitted predictions from the CASP11's website.
If the underlying gene did not have an annotated UTR and a start codon, at least one of the predictors' predictions had to be based on more than the N-terminal region of the protein.
Unlike conventional regression, confidence predictor produces prediction intervals that satisfy a required confidence level.
With the proposed predictor, the prediction errors can be efficiently reduced for different types of audio files.
We also found support for temporal effects, particularly in tests that included estrous overlap as a predictor variable (Prediction iii).
To evaluate the performance of the predictor, the prediction accuracy, specificity, sensitivity, and MCC (Matthews's correlation coefficient) were calculated as follows: (1) accuracy = TP + TN TP + TN + FP + FN, sensitivity = TP TP + FN, specificity = TN TN + FP, MCC = TP × TN − FP × FN (TP + FP ) (TP + FN ) (TN + FP ) (TN + FN ), where TP denotes true positive.
When it was removed from the predictors, the prediction accuracy strongly decreased.
For all predictors, the prediction scores were used as returned by the web server.
