eval: Accuracy measures for quantitative or categorical variables

In AlexandreWadoux/soilspec: The soilspec R package

Description

Computation of a set of accuracy measures between observed and predicted values or classes.

Usage

eval(obs, pred, obj, ...)

Arguments

obs	A vector of measured values or classes.
pred	A vector of predicted values or classes.
obj	Either quant or cat to specify whether obs and pred are quantitative or categorical.
...	Further not yet used parameters.

Details

The accuracy measures for quantitative variables (when obj = 'quant') are

• ME: mean error

• RMSE: root mean square error

• r2: squared Pearson correlation coefficient

• R2: coefficient of determination

• rhoC: concordance correlation coefficient

• RPD: ratio of performance to deviation

• RPIQ: ratio of performance to inter-quartile distance

The accuracy measures for categorical variables (when obj = 'cat'):

• OA: overall accuracy

• kappa: Cohen's kappa statistic

Value

Data frame with accuracy measures.

AlexandreWadoux/soilspec documentation built on March 7, 2021, 1:54 p.m.