ComparisonMetrics (FSharp.Stats)

ComparisonMetrics Type

Namespace: FSharp.Stats.Testing
Assembly: FSharp.Stats.dll
Base Type: obj
All Interfaces: IEquatable<ComparisonMetrics> , IStructuralEquatable , IComparable<ComparisonMetrics> , IComparable , IStructuralComparable

Comparison metrics that can be derived from a binary confusion matrix

Record fields

Record Field	Description
`Accuracy` Full Usage: `Accuracy` Field type: `float`	accuracy (ACC) Field type: `float`
`BalancedAccuracy` Full Usage: `BalancedAccuracy` Field type: `float`	balanced accuracy (BA) Field type: `float`
`DiagnosticOddsRatio` Full Usage: `DiagnosticOddsRatio` Field type: `float`	Diagnostic odds ratio (DOR) Field type: `float`
`F1` Full Usage: `F1` Field type: `float`	F1 score (harmonic mean of precision and sensitivity) Field type: `float`
`FN` Full Usage: `FN` Field type: `float`	false negatives: condition-positive labels incorrectly predicted as negatives Field type: `float`
`FP` Full Usage: `FP` Field type: `float`	false positives: condition-negative labels incorrectly predicted as positives Field type: `float`
`FallOut` Full Usage: `FallOut` Field type: `float`	fall-out or false positive rate (FPR) Field type: `float`
`FalseDiscoveryRate` Full Usage: `FalseDiscoveryRate` Field type: `float`	false discovery rate (FDR) Field type: `float`
`FalseOmissionRate` Full Usage: `FalseOmissionRate` Field type: `float`	false omission rate (FOR) Field type: `float`
`FowlkesMallowsIndex` Full Usage: `FowlkesMallowsIndex` Field type: `float`	Fowlkes–Mallows index (FM) Field type: `float`
`Informedness` Full Usage: `Informedness` Field type: `float`	informedness or bookmaker informedness (BM) Field type: `float`
`Markedness` Full Usage: `Markedness` Field type: `float`	markedness (MK) or deltaP (Δp) Field type: `float`
`Missrate` Full Usage: `Missrate` Field type: `float`	miss rate or false negative rate (FNR) Field type: `float`
`N` Full Usage: `N` Field type: `float`	Condition-negatives: all negative labels in the sample Field type: `float`
`NegativeLikelihoodRatio` Full Usage: `NegativeLikelihoodRatio` Field type: `float`	Negative likelihood ratio (LR-) Field type: `float`
`NegativePredictiveValue` Full Usage: `NegativePredictiveValue` Field type: `float`	negative predictive value (NPV) Field type: `float`
`P` Full Usage: `P` Field type: `float`	Condition-positives: all positive labels in the sample Field type: `float`
`PhiCoefficient` Full Usage: `PhiCoefficient` Field type: `float`	phi coefficient (φ or rφ) or Matthews correlation coefficient (MCC) Field type: `float`
`PositiveLikelihoodRatio` Full Usage: `PositiveLikelihoodRatio` Field type: `float`	Positive likelihood ratio (LR+) Field type: `float`
`Precision` Full Usage: `Precision` Field type: `float`	precision or positive predictive value (PPV) Field type: `float`
`Prevalence` Full Usage: `Prevalence` Field type: `float`	Prevalence Field type: `float`
`PrevalenceThreshold` Full Usage: `PrevalenceThreshold` Field type: `float`	prevalence threshold (PT) Field type: `float`
`SampleSize` Full Usage: `SampleSize` Field type: `float`	all observations in the sample Field type: `float`
`Sensitivity` Full Usage: `Sensitivity` Field type: `float`	sensitivity, recall, hit rate, or true positive rate (TPR) Field type: `float`
`Specificity` Full Usage: `Specificity` Field type: `float`	specificity, selectivity or true negative rate (TNR) Field type: `float`
`TN` Full Usage: `TN` Field type: `float`	true negatives: correctly predicted condition-negative labels Field type: `float`
`TP` Full Usage: `TP` Field type: `float`	true positives: correctly predicted condition-positive labels Field type: `float`
`ThreatScore` Full Usage: `ThreatScore` Field type: `float`	threat score (TS) or critical success index (CSI) Field type: `float`

Static members

Static member	Description
`ComparisonMetrics.binaryThresholdMap (actual, predictions)` Full Usage: `ComparisonMetrics.binaryThresholdMap (actual, predictions)` Parameters: actual : `seq<bool>` predictions : `seq<float>` Returns: `(float * ComparisonMetrics)[]`	actual : `seq<bool>` predictions : `seq<float>` Returns: `(float * ComparisonMetrics)[]`
`ComparisonMetrics.binaryThresholdMap (actual, predictions, thresholds)` Full Usage: `ComparisonMetrics.binaryThresholdMap (actual, predictions, thresholds)` Parameters: actual : `seq<bool>` predictions : `seq<float>` thresholds : `seq<float>` Returns: `(float * ComparisonMetrics)[]`	actual : `seq<bool>` predictions : `seq<float>` thresholds : `seq<float>` Returns: `(float * ComparisonMetrics)[]`
`ComparisonMetrics.binaryThresholdMap tm` Full Usage: `ComparisonMetrics.binaryThresholdMap tm` Parameters: tm : `(float * BinaryConfusionMatrix)[]` Returns: `(float * ComparisonMetrics)[]`	tm : `(float * BinaryConfusionMatrix)[]` Returns: `(float * ComparisonMetrics)[]`
`ComparisonMetrics.calculateAccuracy tp tn samplesize` Full Usage: `ComparisonMetrics.calculateAccuracy tp tn samplesize` Parameters: tp : `float` tn : `float` samplesize : `float` Returns: `float`	calculates the accuracy (ACC) tp : `float` tn : `float` samplesize : `float` Returns: `float`
`ComparisonMetrics.calculateBalancedAccuracy tp p tn n` Full Usage: `ComparisonMetrics.calculateBalancedAccuracy tp p tn n` Parameters: tp : `float` p : `float` tn : `float` n : `float` Returns: `float`	calculates the balanced accuracy (BA) tp : `float` p : `float` tn : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateDiagnosticOddsRatio tp tn fp fn p n` Full Usage: `ComparisonMetrics.calculateDiagnosticOddsRatio tp tn fp fn p n` Parameters: tp : `float` tn : `float` fp : `float` fn : `float` p : `float` n : `float` Returns: `float`	calculates the Diagnostic odds ratio (DOR) tp : `float` tn : `float` fp : `float` fn : `float` p : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateF1 tp fp fn` Full Usage: `ComparisonMetrics.calculateF1 tp fp fn` Parameters: tp : `float` fp : `float` fn : `float` Returns: `float`	calculates the F1 score (harmonic mean of precision and sensitivity) tp : `float` fp : `float` fn : `float` Returns: `float`
`ComparisonMetrics.calculateFallOut fp n` Full Usage: `ComparisonMetrics.calculateFallOut fp n` Parameters: fp : `float` n : `float` Returns: `float`	calculates the fall-out or false positive rate (FPR) fp : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateFalseDiscoveryRate fp tp` Full Usage: `ComparisonMetrics.calculateFalseDiscoveryRate fp tp` Parameters: fp : `float` tp : `float` Returns: `float`	calculates the false discovery rate (FDR) fp : `float` tp : `float` Returns: `float`
`ComparisonMetrics.calculateFalseOmissionRate fn tn` Full Usage: `ComparisonMetrics.calculateFalseOmissionRate fn tn` Parameters: fn : `float` tn : `float` Returns: `float`	calculates the false omission rate (FOR) fn : `float` tn : `float` Returns: `float`
`ComparisonMetrics.calculateFowlkesMallowsIndex tp fp p` Full Usage: `ComparisonMetrics.calculateFowlkesMallowsIndex tp fp p` Parameters: tp : `float` fp : `float` p : `float` Returns: `float`	calculates the Fowlkes–Mallows index (FM) tp : `float` fp : `float` p : `float` Returns: `float`
`ComparisonMetrics.calculateInformedness tp p tn n` Full Usage: `ComparisonMetrics.calculateInformedness tp p tn n` Parameters: tp : `float` p : `float` tn : `float` n : `float` Returns: `float`	calculates the informedness or bookmaker informedness (BM) tp : `float` p : `float` tn : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateMarkedness tp fp tn fn` Full Usage: `ComparisonMetrics.calculateMarkedness tp fp tn fn` Parameters: tp : `float` fp : `float` tn : `float` fn : `float` Returns: `float`	calculates the markedness (MK) or deltaP (Δp) tp : `float` fp : `float` tn : `float` fn : `float` Returns: `float`
`ComparisonMetrics.calculateMissrate fn p` Full Usage: `ComparisonMetrics.calculateMissrate fn p` Parameters: fn : `float` p : `float` Returns: `float`	calculates the miss rate or false negative rate (FNR) fn : `float` p : `float` Returns: `float`
`ComparisonMetrics.calculateMultiLabelROC (actual, predictions)` Full Usage: `ComparisonMetrics.calculateMultiLabelROC (actual, predictions)` Parameters: actual : `'a[]` predictions : `('a * float[])[]` Returns: `Map<string, (float * float)[]>`	actual : `'a[]` predictions : `('a * float[])[]` Returns: `Map<string, (float * float)[]>`
`ComparisonMetrics.calculateNegativeLikelihoodRatio fn p tn n` Full Usage: `ComparisonMetrics.calculateNegativeLikelihoodRatio fn p tn n` Parameters: fn : `float` p : `float` tn : `float` n : `float` Returns: `float`	calculates the Negative likelihood ratio (LR-) fn : `float` p : `float` tn : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateNegativePredictiveValue tn fn` Full Usage: `ComparisonMetrics.calculateNegativePredictiveValue tn fn` Parameters: tn : `float` fn : `float` Returns: `float`	calculates the negative predictive value (NPV) tn : `float` fn : `float` Returns: `float`
`ComparisonMetrics.calculatePhiCoefficient tp tn fp fn` Full Usage: `ComparisonMetrics.calculatePhiCoefficient tp tn fp fn` Parameters: tp : `float` tn : `float` fp : `float` fn : `float` Returns: `float`	calculates the phi coefficient (φ or rφ) or Matthews correlation coefficient (MCC) tp : `float` tn : `float` fp : `float` fn : `float` Returns: `float`
`ComparisonMetrics.calculatePositiveLikelihoodRatio tp p fp n` Full Usage: `ComparisonMetrics.calculatePositiveLikelihoodRatio tp p fp n` Parameters: tp : `float` p : `float` fp : `float` n : `float` Returns: `float`	calculates the Positive likelihood ratio (LR+) tp : `float` p : `float` fp : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculatePrecision tp fp` Full Usage: `ComparisonMetrics.calculatePrecision tp fp` Parameters: tp : `float` fp : `float` Returns: `float`	calculates the precision or positive predictive value (PPV) tp : `float` fp : `float` Returns: `float`
`ComparisonMetrics.calculatePrevalence p samplesize` Full Usage: `ComparisonMetrics.calculatePrevalence p samplesize` Parameters: p : `float` samplesize : `float` Returns: `float`	calculates the Prevalence p : `float` samplesize : `float` Returns: `float`
`ComparisonMetrics.calculatePrevalenceThreshold fp n tp p` Full Usage: `ComparisonMetrics.calculatePrevalenceThreshold fp n tp p` Parameters: fp : `float` n : `float` tp : `float` p : `float` Returns: `float`	calculates the prevalence threshold (PT) fp : `float` n : `float` tp : `float` p : `float` Returns: `float`
`ComparisonMetrics.calculateROC (actual, predictions)` Full Usage: `ComparisonMetrics.calculateROC (actual, predictions)` Parameters: actual : `seq<bool>` predictions : `seq<float>` Returns: `(float * float)[]`	actual : `seq<bool>` predictions : `seq<float>` Returns: `(float * float)[]`
`ComparisonMetrics.calculateROC (actual, predictions, thresholds)` Full Usage: `ComparisonMetrics.calculateROC (actual, predictions, thresholds)` Parameters: actual : `seq<bool>` predictions : `seq<float>` thresholds : `seq<float>` Returns: `(float * float)[]`	actual : `seq<bool>` predictions : `seq<float>` thresholds : `seq<float>` Returns: `(float * float)[]`
`ComparisonMetrics.calculateSensitivity tp p` Full Usage: `ComparisonMetrics.calculateSensitivity tp p` Parameters: tp : `float` p : `float` Returns: `float`	calculates the sensitivity, recall, hit rate, or true positive rate (TPR) tp : `float` p : `float` Returns: `float`
`ComparisonMetrics.calculateSpecificity tn n` Full Usage: `ComparisonMetrics.calculateSpecificity tn n` Parameters: tn : `float` n : `float` Returns: `float`	calculates the specificity, selectivity or true negative rate (TNR) tn : `float` n : `float` Returns: `float`
`ComparisonMetrics.calculateThreatScore tp fn fp` Full Usage: `ComparisonMetrics.calculateThreatScore tp fn fp` Parameters: tp : `float` fn : `float` fp : `float` Returns: `float`	calculates the threat score (TS) or critical success index (CSI) tp : `float` fn : `float` fp : `float` Returns: `float`
`ComparisonMetrics.create bcm` Full Usage: `ComparisonMetrics.create bcm` Parameters: bcm : `BinaryConfusionMatrix` Returns: `ComparisonMetrics`	bcm : `BinaryConfusionMatrix` Returns: `ComparisonMetrics`
`ComparisonMetrics.create (tp, tn, fp, fn)` Full Usage: `ComparisonMetrics.create (tp, tn, fp, fn)` Parameters: tp : `float` tn : `float` fp : `float` fn : `float` Returns: `ComparisonMetrics`	tp : `float` tn : `float` fp : `float` fn : `float` Returns: `ComparisonMetrics`
`ComparisonMetrics.create (p, n, samplesize, tp, tn, fp, fn, sensitivity, specificity, precision, negativepredictivevalue, missrate, fallout, falsediscoveryrate, falseomissionrate, positivelikelihoodratio, negativelikelihoodratio, prevalencethreshold, threatscore, prevalence, accuracy, balancedaccuracy, f1, phicoefficient, fowlkesmallowsindex, informedness, markedness, diagnosticoddsratio)` Full Usage: `ComparisonMetrics.create (p, n, samplesize, tp, tn, fp, fn, sensitivity, specificity, precision, negativepredictivevalue, missrate, fallout, falsediscoveryrate, falseomissionrate, positivelikelihoodratio, negativelikelihoodratio, prevalencethreshold, threatscore, prevalence, accuracy, balancedaccuracy, f1, phicoefficient, fowlkesmallowsindex, informedness, markedness, diagnosticoddsratio)` Parameters: p : `float` n : `float` samplesize : `float` tp : `float` tn : `float` fp : `float` fn : `float` sensitivity : `float` specificity : `float` precision : `float` negativepredictivevalue : `float` missrate : `float` fallout : `float` falsediscoveryrate : `float` falseomissionrate : `float` positivelikelihoodratio : `float` negativelikelihoodratio : `float` prevalencethreshold : `float` threatscore : `float` prevalence : `float` accuracy : `float` balancedaccuracy : `float` f1 : `float` phicoefficient : `float` fowlkesmallowsindex : `float` informedness : `float` markedness : `float` diagnosticoddsratio : `float` Returns: `ComparisonMetrics`	p : `float` n : `float` samplesize : `float` tp : `float` tn : `float` fp : `float` fn : `float` sensitivity : `float` specificity : `float` precision : `float` negativepredictivevalue : `float` missrate : `float` fallout : `float` falsediscoveryrate : `float` falseomissionrate : `float` positivelikelihoodratio : `float` negativelikelihoodratio : `float` prevalencethreshold : `float` threatscore : `float` prevalence : `float` accuracy : `float` balancedaccuracy : `float` f1 : `float` phicoefficient : `float` fowlkesmallowsindex : `float` informedness : `float` markedness : `float` diagnosticoddsratio : `float` Returns: `ComparisonMetrics`
`ComparisonMetrics.macroAverage bcms` Full Usage: `ComparisonMetrics.macroAverage bcms` Parameters: bcms : `BinaryConfusionMatrix[]` Returns: `ComparisonMetrics`	bcms : `BinaryConfusionMatrix[]` Returns: `ComparisonMetrics`
`ComparisonMetrics.macroAverage mlcm` Full Usage: `ComparisonMetrics.macroAverage mlcm` Parameters: mlcm : `MultiLabelConfusionMatrix` Returns: `ComparisonMetrics`	mlcm : `MultiLabelConfusionMatrix` Returns: `ComparisonMetrics`
`ComparisonMetrics.macroAverage metrics` Full Usage: `ComparisonMetrics.macroAverage metrics` Parameters: metrics : `seq<ComparisonMetrics>` Returns: `ComparisonMetrics`	calculates comparison metrics as macro average of the given sequence of comparison metrics (all metrics are calculated as the average of the respective metrics) metrics : `seq<ComparisonMetrics>` Returns: `ComparisonMetrics`
`ComparisonMetrics.macroAverageOfMultiLabelPredictions (labels, actual, predictions)` Full Usage: `ComparisonMetrics.macroAverageOfMultiLabelPredictions (labels, actual, predictions)` Parameters: labels : `'a[]` actual : `'a[]` predictions : `'a[]` Returns: `ComparisonMetrics`	labels : `'a[]` actual : `'a[]` predictions : `'a[]` Returns: `ComparisonMetrics`
`ComparisonMetrics.microAverage mlcm` Full Usage: `ComparisonMetrics.microAverage mlcm` Parameters: mlcm : `MultiLabelConfusionMatrix` Returns: `ComparisonMetrics`	calculates comparison metrics from the given MultiLabelConfusionMatrix as micro-averages (one-vs-rest binary confusion matrices (TP/TN/FP/FN) are calculated for each label and then aggregated before calculating metrics) mlcm : `MultiLabelConfusionMatrix` Returns: `ComparisonMetrics`
`ComparisonMetrics.microAverage cms` Full Usage: `ComparisonMetrics.microAverage cms` Parameters: cms : `seq<BinaryConfusionMatrix>` Returns: `ComparisonMetrics`	calculates comparison metrics from multiple binary confusion matrices as micro-averages (all TP/TN/FP/FN are aggregated before calculating metrics) cms : `seq<BinaryConfusionMatrix>` Returns: `ComparisonMetrics`
`ComparisonMetrics.microAverageOfMultiLabelPredictions (labels, actual, predictions)` Full Usage: `ComparisonMetrics.microAverageOfMultiLabelPredictions (labels, actual, predictions)` Parameters: labels : `'a[]` actual : `'a[]` predictions : `'a[]` Returns: `ComparisonMetrics`	labels : `'a[]` actual : `'a[]` predictions : `'a[]` Returns: `ComparisonMetrics`
`ComparisonMetrics.multiLabelThresholdMap (actual, predictions)` Full Usage: `ComparisonMetrics.multiLabelThresholdMap (actual, predictions)` Parameters: actual : `'a[]` predictions : `('a * float[])[]` Returns: `Map<string, (float * ComparisonMetrics)[]>`	actual : `'a[]` predictions : `('a * float[])[]` Returns: `Map<string, (float * ComparisonMetrics)[]>`
`ComparisonMetrics.ofBinaryPredictions (actual, predictions)` Full Usage: `ComparisonMetrics.ofBinaryPredictions (actual, predictions)` Parameters: actual : `seq<bool>` predictions : `seq<bool>` Returns: `ComparisonMetrics`	actual : `seq<bool>` predictions : `seq<bool>` Returns: `ComparisonMetrics`
`ComparisonMetrics.ofBinaryPredictions (positiveLabel, actual, predictions)` Full Usage: `ComparisonMetrics.ofBinaryPredictions (positiveLabel, actual, predictions)` Parameters: positiveLabel : `'A` actual : `seq<'A>` predictions : `seq<'A>` Returns: `ComparisonMetrics`	positiveLabel : `'A` actual : `seq<'A>` predictions : `seq<'A>` Returns: `ComparisonMetrics`