score: Score Function for Metrics

In stressor: Algorithms for Testing Models under Stress

Score Function for Metrics

Description

A score function takes the observed and predicted values and returns a vector or data.frame of the various metrics that are reported from 'PyCaret'. For regression, the following metrics are available: 'RMSE', 'MAE', 'MSE', 'R2', 'RMSLE', and 'MAPE'. For classification, the following metrics are available:'Accuracy', 'AUC', 'Recall', 'Prec.', 'F1', 'MCC', and 'Kappa'.

Usage

score(observed, predicted, ...)

Arguments

observed	A vector of the observed results.
predicted	A data.frame or vector object that is the same number of rows or length as the length of observed values.
...	Arguments passed on to score_classification, score_regression metrics A character vector of the metrics to be fitted. This is defaulted to be the metrics from 'PyCaret'.

Value

A matrix with the various metrics reported.

Examples

lm_data <- data_gen_lm(100)
indices <- split_data_prob(lm_data, .2)
train <- lm_data[!indices,]
test <- lm_data[indices,]
model <- lm(Y ~ ., train)
pred_lm <- predict(model, test)
score(test$Y, pred_lm)

stressor documentation built on May 29, 2024, 3:09 a.m.