Struct tangram_metrics::BinaryClassificationMetricsOutputForThreshold [−][src]
pub struct BinaryClassificationMetricsOutputForThreshold { pub threshold: f32, pub true_positives: u64, pub false_positives: u64, pub true_negatives: u64, pub false_negatives: u64, pub accuracy: f32, pub precision: Option<f32>, pub recall: Option<f32>, pub f1_score: Option<f32>, pub true_positive_rate: f32, pub false_positive_rate: f32, }
Expand description
The output from BinaryClassificationMetrics
.
Fields
threshold: f32
The classification threshold.
true_positives: u64
The total number of examples whose label is equal to the positive class that the model predicted as belonging to the positive class.
false_positives: u64
The total number of examples whose label is equal to the negative class that the model predicted as belonging to the positive class.
true_negatives: u64
The total number of examples whose label is equal to the negative class that the model predicted as belonging to the negative class.
false_negatives: u64
The total number of examples whose label is equal to the positive class that the model predicted as belonging to the negative class.
accuracy: f32
The fraction of examples that were correctly classified.
precision: Option<f32>
The precision is the fraction of examples the model predicted as belonging to the positive class whose label is actually the positive class. true_positives / (true_positives + false_positives). See Precision and Recall.
recall: Option<f32>
The recall is the fraction of examples whose label is equal to the positive class that the model predicted as belonging to the positive class. recall = true_positives / (true_positives + false_negatives)
.
f1_score: Option<f32>
The f1 score is the harmonic mean of the precision and the recall. See F1 Score.
true_positive_rate: f32
The true positive rate is the fraction of examples whose label is equal to the positive class that the model predicted as belonging to the positive class. Also known as the recall. See Sensitivity and Specificity.
false_positive_rate: f32
The false positive rate is the fraction of examples whose label is equal to the negative class that the model falsely predicted as belonging to the positive class. false_positives / (false_positives + true_negatives). See False Positive Rate
Trait Implementations
Auto Trait Implementations
Blanket Implementations
Mutably borrows from an owned value. Read more