Table 2 Performance statistics for neural networks trained on valence and context of production.

From: Classification of pig calls produced from birth to slaughter according to their emotional valence and context of production

 

Valence

Context

Accuracy

0.915 ± 0.003

0.815 ± 0.003

Precision

0.919 ± 0.005

0.815 ± 0.003

Recall

0.912 ± 0.003

0.813 ± 0.003

F1 score

0.916 ± 0.003

0.812 ± 0.003

  1. For the binary valence classifier (2 classes: positive and negative), the following statistics were computed using the binary precision, recall, and F1 score formulas while treating positive valence labels as positive. For the imbalanced multi-class context classifier (19 classes, Supplementary Table S1), the following statistics were calculated as weighted averages across the classes. From the 10 trials, the mean accuracy, precision, recall, and F1 scores of the classifiers are listed. The uncertainty value is calculated across 10 trials.
Back to article page