Struct tangram_metrics::MulticlassClassificationMetricsOutput [−][src]
pub struct MulticlassClassificationMetricsOutput { pub class_metrics: Vec<ClassMetrics>, pub accuracy: f32, pub precision_unweighted: f32, pub precision_weighted: f32, pub recall_unweighted: f32, pub recall_weighted: f32, }
Expand description
The output from MulticlassClassificationMetrics
.
Fields
class_metrics: Vec<ClassMetrics>
The class metrics contain class specific metrics.
accuracy: f32
The accuracy is the fraction of all of the predictions that are correct.
precision_unweighted: f32
The unweighted precision equal to the mean of each class’s precision.
precision_weighted: f32
The weighted precision is a weighted mean of each class’s precision weighted by the fraction of the total examples in the class.
recall_unweighted: f32
The unweighted recall equal to the mean of each class’s recall.
recall_weighted: f32
The weighted recall is a weighted mean of each class’s recall weighted by the fraction of the total examples in the class.