regressionStats: calculate prediction performance statistics of regression...

In HannaMeyer/sptm: Supporting functions for spatio-temporal models using machine learning

Description

this function calculates prediction performance statistics between vectors of predicted and observed values, namely coefficient of determination (Rsq), root mean squared error (RMSE), mean error (ME), mean absolute error (MAE).

Usage

regressionStats(prd, obs, adj.rsq = TRUE, method = "pearson")

Arguments

prd	numeric vector of predicted values
obs	numeric vector of observed values
adj.rsq	logical, whether to return adjusted r-squared. Defaults to TRUE
method	character. Method to use for correlation. See ?cor.test for details.

Value

a data frame.

Author(s)

Tim Appelhans, Hanna Meyer

Examples

## create predictions with high accuracy (identical mean),
## but low precision (sd double that of observations). Hence,
## ME should be close to zero and RMSE close to ten.
pred_vals <- sort(rnorm(1000, 200, 20)) # sorting ensures high Rsq
obs_vals <- sort(rnorm(1000, 200, 10))
result <- regressionStats(pred_vals, obs_vals, adj.rsq = FALSE)
result

HannaMeyer/sptm documentation built on May 15, 2019, 4:29 p.m.