kappa_multinomial: kappa_multinomial
In bobdouma/kappa_multinomial: Evaluate the performance of probabilistic multinomial models

Description Usage Arguments Details Value Author(s) References See Also Examples

A kappa-like measure for multinomial models

1	kappa_multinomial(obs, pred, nsim = 1000)

`obs`	a data.frame with class obervations with n columns and m rows; each row represents a sample, the columns represent the classes of outcomes.
`pred`	a data.frame with class predictions with n columns and m rows; each rows represents a sample, the columns represent the classes of outcomes.
`nsim`	to be passed to the null model if null model is computed based on randmization. nsim represents the number of randomizations

kappa_multinomial is caclulated from two components: kappa_loc and kappa_prob. For details see Douma et al 2017.

obs and pred should contain probabilities or discrete outcomes (0 or 1).

The null model is calculated analytically when observations are discrete and calculated by randomzation otherwise.

returns a list with the following elements:

Kappa_prob
Kappa_loc
Kappa_multinomial
p0
pe
pmax

Bob Douma

Douma et al 2017 Journal of Biogeography

cohen.kappa for Cohen's Kappa

# generate multinomial probabilties with four classes
pred = data.frame(gtools::rdirichlet(100, c(0.1,0.1,0.5,0.5)))
# generate multinomial observations with four classes 
obs = data.frame(t(apply(pred,1,rmultinom,size=1,n=1)))
# calculate kappa multinomial 
kappa_multinomial(obs=obs,pred=pred)