metric_Fbeta: metric_Fbeta

In gloverd2/admr: Contains standard functions for Admiral Advance Analytics

Description

Returns the Fbeta [(1 + beta^2) * (precision * recall) / ((beta^2 *precision) + recall)] of a classification using the confusion matrix Note: Predictions should be annualized (independent of exposure) Note: Perfect Fbeta is 1, poor model is 0

Usage

metric_Fbeta(
  actual,
  predicted,
  weight = NULL,
  na.rm = FALSE,
  threshold = 0.5,
  beta = 1
)

Arguments

actual	Array[Numeric] - Values we are aiming to predict.
predicted	Array[Numeric] - Values that we have predicted.
weight	Optional: Array[Numeric] - Weighting of predictions. If NULL even weighting is used
na.rm	Optional: boolean - If FALSE function will return NA is any value in NA
threshold	Optional: Numeric between 0 and 1. If prediction proablity is below threshold the predicted value is 0.
beta	Numeric

Value

precision of classification TP / (TP + FN)

See Also

metric_precision, metric_recall and metric_confusion_matrix

Examples

metric_Fbeta(actual=c(0,1,0,0), predicted=c(0.1,0.9,0.4,0.6))
metric_Fbeta(actual=c(0,1,0,0), predicted=c(0.1,0.9,0.4,0.6), threshold=0.7)

gloverd2/admr documentation built on Dec. 2, 2020, 11:16 p.m.