ctmaInit: ctmaInit

In CoTiMA: Continuous Time Meta-Analysis ('CoTiMA')

ctmaInit

Description

Fits ctsem models to each primary study in the supplied list of primary studies prepared by ctmaPrep.

Usage

ctmaInit(
  activateRPB = FALSE,
  activeDirectory = NULL,
  binaries = NULL,
  chains = NULL,
  checkSingleStudyResults = FALSE,
  cint = 0,
  coresToUse = c(2),
  CoTiMAStanctArgs = NULL,
  customPar = FALSE,
  diff = NULL,
  digits = 4,
  doPar = 1,
  drift = NULL,
  experimental = FALSE,
  finishsamples = NULL,
  fit = TRUE,
  indVarying = FALSE,
  indVaryingT0 = NULL,
  iter = NULL,
  lambda = NULL,
  loadSingleStudyModelFit = c(),
  manifestMeans = 0,
  manifestVars = NULL,
  n.latent = NULL,
  n.manifest = 0,
  optimize = TRUE,
  primaryStudies = NULL,
  priors = FALSE,
  randomIntercepts = FALSE,
  sameInitialTimes = FALSE,
  saveRawData = list(),
  saveSingleStudyModelFit = c(),
  scaleTI = NULL,
  scaleTime = NULL,
  silentOverwrite = FALSE,
  T0means = 0,
  T0var = "auto",
  useSV = FALSE,
  verbose = 0
)

Arguments

activateRPB	set to TRUE to receive push messages with 'CoTiMA' notifications on your phone
activeDirectory	defines another active directory than the one used in ctmaPrep
binaries	which manifest is a binary. Still experimental
chains	number of chains to sample, during HMC or post-optimization importance sampling.
checkSingleStudyResults	Displays estimates from single study ctsem models and waits for user input to continue. Useful to check estimates before they are saved.
cint	default 'auto' (= 0). Are set free if random intercepts model with varying cints is requested (by indVarying='cint')
coresToUse	if neg., the value is subtracted from available cores, else value = cores to use
CoTiMAStanctArgs	parameters that can be set to improve model fitting of the ctStanFit Function
customPar	logical. If set TRUE leverages the first pass using priors and ensure that the drift diagonal cannot easily go too negative (helps since ctsem > 3.4)
diff	labels for diffusion effects. Have to be either of the character strings of the type "diff_eta1" or "diff_eta2_eta1" (= freely estimated) or values (e.g., 0 for effects to be excluded, which is usually not recommended)
digits	number of digits used for rounding (in outputs)
doPar	parallel and multiple fitting if single studies. A value > 1 will fit each study doPar times in parallel mode during which no output is generated (screen remains silent). Useful to obtain best fit.
drift	labels for drift effects. Have to be either of the character strings of the type V1toV2 (= freely estimated) or values (e.g., 0 for effects to be excluded, which is usually not recommended)
experimental	used for debugging puposes (default = FALSE)
finishsamples	number of samples to draw (either from hessian based covariance or posterior distribution) for final results computation (default = 1000).
fit	TRUE (default) fits the requested model. FALSE returns the ctsem code CoTiMA uses to set up the model, the ctsemmodelbase which can be modified to match users requirements, and the data set (in long format created). The model can then be fitted using ctStanFit)
indVarying	control for unobserved heterogeneity by having randomly (inter-individually) varying manifest means
indVaryingT0	deprecated. Automatically set to NULL.
iter	number of interation (defaul = 1000). Sometimes larger values could be required fom Bayesian estimation
lambda	R-type matrix with pattern of fixed (=1) or free (any string) loadings.
loadSingleStudyModelFit	load the fit of single study ctsem models
manifestMeans	Default 0 (assuming standardized variables). Can be assigned labels to estimate them freely.
manifestVars	define the error variances of the manifests within a single time point using R-type lower triangular matrix with nrow=n.manifest & ncol=n.manifest.
n.latent	number of latent variables of the model (hast to be specified)!
n.manifest	number of manifest variables of the model (if left empty it will assumed to be identical with n.latent).
optimize	if set to FALSE, Stan's Hamiltonian Monte Carlo sampler is used (default = TRUE = maximum a posteriori / importance sampling) .
primaryStudies	list of primary study information created with ctmaPrep
priors	if FALSE, any priors are disabled – sometimes desirable for optimization
randomIntercepts	(default = FALSE) Experimental. Overrides ctsem's default mode for modelling indVarying cints.
sameInitialTimes	Only important for raw data. If TRUE (default=FALSE), T0MEANS occurs for every subject at the same time, rather than just at the earliest observation.
saveRawData	save (created pseudo) raw date. List: saveRawData$studyNumbers, $fileName, $row.names, col.names, $sep, $dec
saveSingleStudyModelFit	save the fit of single study ctsem models (could save a lot of time afterwards if the fit is loaded)
scaleTI	scale TI predictors
scaleTime	scale time (interval) - sometimes desirable to improve fitting
silentOverwrite	overwrite old files without asking
T0means	Default 0 (assuming standardized variables). Can be assigned labels to estimate them freely.
T0var	(default = 'auto')
useSV	if TRUE (default=FALSE) start values will be used if provided in the list of primary studies
verbose	integer from 0 to 2. Higher values print more information during model fit - for debugging

Value

ctmaFit returns a list containing some arguments supplied, the fitted models, different elements summarizing the main results, model type, and the type of plot that could be performed with the returned object. The arguments in the returned object are activeDirectory, coresToUse, n.latent, n.manifest, and primaryStudyList. The study count is returned as n.studies, the created matrix of loadings of manifest on latent factors is returned as lambda, and a re-organized list of primary studies with some information ommited is returned as studyList. The fitted models for each primary study are found in studyFitList, which is a large list with many elements (e.g., the ctsem model specified by CoTiMA, the rstan model created by ctsem, the fitted rstan model etc.). Further results returned are emprawList (containing the pseudo raw data created), statisticsList (comprising baisc stats such as average sample size, no. of measurement points, etc.), a list with modelResults (i.e., DRIFT=model_Drift_Coef, DIFFUSION=model_Diffusion_Coef, T0VAR=model_T0var_Coef, CINT=model_Cint_Coef), and the paramter names internally used. The summary list, which is printed if the summary function is applied to the returned object, comprises "estimates" (the aggregated effects), possible randomIntercepts,confidenceIntervals, the minus2ll value and its n.parameters, and possible warning messages (message). Plot type is plot.type=c("drift") and model.type="stanct" ("omx" was deprecated).

Examples

# Fit a ctsem model to all three primary studies summarized in
# CoTiMAstudyList_3 and save the three fitted models
## Not run: 
CoTiMAInitFit_3 <- ctmaInit(primaryStudies=CoTiMAstudyList_3,
                            n.latent=2,
                            checkSingleStudyResults=FALSE,
                            activeDirectory="/Users/tmp/") # adapt!
summary(CoTiMAInitFit_3)

## End(Not run)

CoTiMA documentation built on May 29, 2024, 11:39 a.m.