jSDM (version 0.1.0)

predict.jSDM: Predict method for models fitted with jSDM

Description

Predicted values for models fitted with jSDM

Usage

# S3 method for jSDM
predict(object, newdata=NULL, Id_species, Id_sites, type="mean",
  probs=c(0.025,0.975), ...)

Arguments

object

An object of class "jSDM".

newdata

An optional data frame in which to look for variables with which to predict. If omitted, the fitted values are used.

Id_species

An vector of character or integer indicating for which species the probabilities of presence on chosen sites will be predicted.

Id_sites

An vector of integer indicating for which sites the probabilities of presence of specified species will be predicted.

type

Type of prediction. Can be "mean" for predictive posterior mean, "quantile" for producing sample quantiles from the predictive posterior corresponding to the given probabilities (see probs argument) or "posterior" for the full predictive posterior for each prediction. Using "quantile" or "posterior" might lead to memory problem depending on the number of predictions and the number of samples for the jSDM model's parameters.

probs

Numeric vector of probabilities with values in [0,1] and used when type="quantile".

...

Further arguments passed to or from other methods.

Value

Return a vector for the predictive posterior mean when type="mean", a data-frame with the mean and quantiles when type="quantile" or an mcmc object (see coda package) with posterior distribution for each prediction when type="posterior".

See Also

jSDM

Examples

# NOT RUN {
# frogs data
data(frogs, package="jSDM")

# Arranging data
PA_frogs <- frogs[,4:12]

# Normalized continuous variables
Env_frogs <- cbind(scale(frogs[,1]),frogs[,2],scale(frogs[,3]))
colnames(Env_frogs) <- colnames(frogs[,1:3])

# Parameter inference
# Increase the number of iterations to reach MCMC convergence
mod_jSDM_block_frogs <- jSDM::jSDM_probit_block(
  # Response variable 
  presence_site_sp = as.matrix(PA_frogs), 
  # Explanatory variables 
  site_suitability = ~.,   
  site_data = as.data.frame(Env_frogs), n_latent=2,
  # Chains
  burnin=1000, mcmc=1000, thin=1,
  # Starting values
  alpha_start=0, beta_start=0,
  lambda_start=0, W_start=0,
  V_alpha_start=1, 
  # Priors
  shape=0.5, rate=0.0005,
  mu_beta=0, V_beta=1.0E6,
  mu_lambda=0, V_lambda=10,
  # Various 
  seed=1234, verbose=1)
  
# Select site and species for predictions
## 30 sites
Id_sites <- sample.int(nrow(PA_frogs), 30)
## 5 species
Id_species <- sample(colnames(PA_frogs), 5)

# Predictions 
theta_pred <- jSDM::predict.jSDM(mod_jSDM_block_frogs,
  Id_species=Id_species, Id_sites=Id_sites, type="mean")
hist(theta_pred, main="Predicted theta with simulated covariates")

# }