The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Initialization algorithms

Cristian Castiglione

Workspace setup

Load the package in the workspace.

library(sgdGMF)

Load other useful packages in the workspace.

library(ggplot2)
library(ggpubr)
library(reshape2)

Ant traits data

Load the ant traits data in the workspace and define the response matrix Y and covariate matrices X and Z.

# install.packages("mvabund")
# data(antTraits, package = "mvabund")

load(url("https://raw.githubusercontent.com/cran/mvabund/master/data/antTraits.RData"))

Y = as.matrix(antTraits$abund)
X = as.matrix(antTraits$env[,-3])
Z = matrix(1, nrow = ncol(Y), ncol = 1)

attr(Y, "dimnames") = NULL
attr(X, "dimnames") = NULL
attr(Z, "dimnames") = NULL

n = nrow(Y)
m = ncol(Y)

Model specification

Set the model family to Poisson since the response matrix contain count data.

family = poisson()
suppressWarnings({
  init_glm_dev = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "glm", type = "deviance")
  init_glm_prs = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "glm", type = "pearson")
  init_glm_lnk = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "glm", type = "link")
  init_ols_dev = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "ols", type = "deviance")
  init_ols_prs = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "ols", type = "pearson")
  init_ols_lnk = sgdGMF::sgdgmf.init(Y, X, Z, ncomp = 2, family = family, method = "ols", type = "link")
})
data.frame(
  "Method" = rep(c("GLM", "OLS"), each = 3),
  "Resid" = rep(c("Deviance", "Pearson", "Link"), times = 2),
  "Deviance" = 
    list(init_glm_dev, init_glm_prs, init_glm_lnk, init_ols_dev, init_ols_prs, init_ols_lnk) |> 
    lapply(function (obj) round(100 * deviance(obj, normalize = TRUE), 2)) |> unlist() |> drop()
)
#>   Method    Resid Deviance
#> 1    GLM Deviance    37.16
#> 2    GLM  Pearson    27.65
#> 3    GLM     Link    45.47
#> 4    OLS Deviance    43.52
#> 5    OLS  Pearson    49.72
#> 6    OLS     Link    42.71

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.