MakeInput_Fn: Build data input for spatial dynamic factor analysis (SDFA)

In James-Thorson/spatial_DFA: Spatial dynamic factor analysis

Description

MakeInput_Fn builds a tagged list of inputs for TMB

Usage

MakeInput_Fn(Version, Nfactors, DF, loc_xy, method = "mesh",
  Nobsfactors = 0, Kappa_Type = "Constant", ObsModel = NULL,
  Aniso = FALSE, Include_Omega = TRUE, Include_Epsilon = TRUE,
  EncounterFunction = 2, Correlated_Overdispersion = FALSE,
  Include_Phi = TRUE, Include_Rho = TRUE, Use_REML = FALSE, X_ik = NULL,
  X_nl = NULL, X_ntl = NULL, a_n = NULL, YearSet = NULL,
  IndependentTF = c(FALSE, FALSE), CheckForBugs = TRUE,
  CorrGroup_pp = NULL, ...)

Arguments

Version	a version number (see example for current default).
Nfactors	The number of dynamic factors used to approximate spatio-temporal variation
DF	a data frame of data where each row is a unique sample and with the following columns catch the observation for each sample year a column of years for each sample spp a factor specifying the species for that sample sitenum a column of years for each sample PredTF a vector of 0s or 1st, stating whether a sample is used when fitting the model (PredTF=0) or when evaluating its predictive performance (PredTF=1) TowID a vector coding the unit of overdispersion, e.g., tow or vessel (OPTIONAL)
loc_xy	Locations for each station
method	The method for approximating spatial variation (options: "grid" or "mesh")
Nobsfactors	The number of factors used to approximate overdispersion among levels of TowID (Default=0)
Kappa_Type	Whether the decorrelation distance is constant for all factors or not (Default="Constant")
ObsModel	The observation model used
Include_Omega	Whether to estimate purely spatial variation (Default=TRUE)
Include_Epsilon,	Whether to incldue spatio-temporal variation (Default=TRUE)
EncounterFunction,	The link between density and encounter probability; 0=two-parameter logistic relationship; 1=two-parameter logistic relationship with less-than-one saturation; 2=one-parameter saturating relationship (Default=1)
Correlated_Overdispersion,	Whether to estimate overdispersion (only possible if TowID is present in DF
Include_Phi	Whether to estimate each factor in equilibrium (FALSE) or with a fixed offset from equilibrium (TRUE), Default=TRUE
Include_Rho	Whether to estimate the magnitude of temporal autocorrelation (Default=TRUE)
Use_REML	Whether to use REML estimation
X_ik	A matrix specifying measured variables that affect catchability for each sample (Default is turned off)
X_nl	A matrix specifying measured variables that affect density for each sample (Default is an intercept)
X_ntl	An array specifying measured variables that affect density and vary over time (Default is off)
a_n	A vector giving the area associated with mesh vertex used when calculating indices of abundance (Default is even weighting of sample locations)
YearSet	A vector of levels of the year input that should be modeled (Default is to infer from the year column of DF)
IndependentTF	Whether the spatio-temporal variation (IndependentTF[1]) or overdispersion (IndependentTF[2]) is independent among species (Default is both are correlated)
CheckForBugs	Whether to check inputs for obvious problems (Default=TRUE)
CorrGroup_pp	A matrix filled with integers, for post hoc testing of differences in among-species correlations

Value

Tagged list containing inputs to function Build_TMB_Fn()

TmbData

A tagged list of data inputs

TmbParams

A tagged list with input parameters

Random

A character vector specifying which parameters are treated as random effects

Map

A parameter map, used to turn off or mirror parameter values

James-Thorson/spatial_DFA documentation built on July 9, 2020, 7:56 a.m.