Skip to main content
. 2016 Mar 11;22:72. doi: 10.1007/s00894-016-2933-0

Fig. 3.

Fig. 3

Relative gains in predictive performance (i.e., in F-measure values and accuracies) achieved upon applying 1-star consensus classifiers based on various types of classifer to CASP-8, CASP-9 and CASP-10 protein targets, as compared to the corresponding average predictive performance of a single classifier of each type for the same targets