opticut-package: Likelihood Based Optimal Partitioning and Indicator Species...

In psolymos/opticut: Likelihood Based Optimal Partitioning and Indicator Species Analysis

Likelihood Based Optimal Partitioning and Indicator Species Analysis

Description

Likelihood based optimal partitioning and indicator species analysis. Finding the best binary partition for each species based on model selection, with the possibility to take into account modifying/confounding variables as described in Kemencei et al. (2014) <doi:10.1556/ComEc.15.2014.2.6>. The package implements binary and multi-level response models, various measures of uncertainty, Lorenz-curve based thresholding, with native support for parallel computations.

Details

The DESCRIPTION file: This package was not yet installed at build time.
Index: This package was not yet installed at build time.

The main user interface are the opticut and multicut functions to find the optimal binary or multi-level response models. Make sure to evaluate uncertainty. optilevels finds the optimal number of factor levels.

Author(s)

Peter Solymos [cre, aut], Ermias T. Azeria [ctb]

Maintainer: Peter Solymos <[e-mail]>

References

Kemencei, Z., Farkas, R., Pall-Gergely, B., Vilisics, F., Nagy, A., Hornung, E. & Solymos, P., 2014. Microhabitat associations of land snails in forested dolinas: implications for coarse filter conservation. Community Ecology 15:180–186. <doi:10.1556/ComEc.15.2014.2.6>

Examples

## community data
y <- cbind(
    Sp1=c(4,6,3,5, 5,6,3,4, 4,1,3,2),
    Sp2=c(0,0,0,0, 1,0,0,1, 4,2,3,4),
    Sp3=c(0,0,3,0, 2,3,0,5, 5,6,3,4))

## stratification
g <-    c(1,1,1,1, 2,2,2,2, 3,3,3,3)

## find optimal partitions for each species
oc <- opticut(formula = y ~ 1, strata = g, dist = "poisson")
summary(oc)

## visualize the results
plot(oc, cut = -Inf)

## quantify uncertainty
uc <- uncertainty(oc, type = "asymp", B = 999)
summary(uc)

## go beyond binary partitions

mc <- multicut(formula = y ~ 1, strata = g, dist = "poisson")
summary(mc)

ol <- optilevels(y[,"Sp2"], as.factor(g))
ol[c("delta", "coef", "rank", "levels")]

psolymos/opticut documentation built on Nov. 27, 2022, 11:29 a.m.