Sentence examples for most alternative classifiers from inspiring English sources

Although most alternative classifiers performed similarly to ENTS on the testing data set, this set included interactions that were used to train the alternatives and could thereby inflate their sensitivities at a given specificity value.

Experimental results show the superiority of the learned features over alternative classifiers constructed using state-of-the-art handcrafted features.

The alternative classifiers differed substantially in their prediction methods.

We found that ENTS predicted more of these interactions than 3 of the 4 alternative classifiers.

We also limited the overlap to training data from the alternative classifiers by choosing datasets containing PPIs that were either published after the alternative classifier's publication date or were not present within the datasets reportedly used for training.

ENTS predictions were then sorted by decreasing confidence and truncated to sets of the same size as the alternative classifiers.

We find that the ratio of experimentally-supported interactions to all positive predictions is higher for ENTS than all alternative classifiers except AtPID.

Some alternative classifier models that can take advantage of such weights can be considered (e.g., bagging of naïve Bayes based classifiers).

On the basis of a simple model, FENSAM provides an alternative classifier for ESCC patients with a high precision.

We use a standard SVM classifier for ranking SMISPs, but also find that simple thresholding rules provide an easily interpretable alternative classifier with accuracy similar to more sophisticated techniques.

First, as the most typical sortal classifier, individual classifiers typically select common nouns, which of course are endurant entities.