metrica - Prediction Performance Metrics
A compilation of more than 80 functions designed to
quantitatively and visually evaluate prediction performance of
regression (continuous variables) and classification
(categorical variables) of point-forecast models (e.g. APSIM,
DSSAT, DNDC, supervised Machine Learning). For regression, it
includes functions to generate plots (scatter, tiles, density,
& Bland-Altman plot), and to estimate error metrics (e.g. MBE,
MAE, RMSE), error decomposition (e.g. lack of
accuracy-precision), model efficiency (e.g. NSE, E1, KGE),
indices of agreement (e.g. d, RAC), goodness of fit (e.g. r,
R2), adjusted correlation coefficients (e.g. CCC, dcorr),
symmetric regression coefficients (intercept, slope), and mean
absolute scaled error (MASE) for time series predictions. For
classification (binomial and multinomial), it offers functions
to generate and plot confusion matrices, and to estimate
performance metrics such as accuracy, precision, recall,
specificity, F-score, Cohen's Kappa, G-mean, and many more. For
more details visit the vignettes
<https://adriancorrendo.github.io/metrica/>.
