grt_wind_fit_parallel: Fit a GRT-wIND model to data several times in parallel

In fsotoc/grtools: General Recognition Theory tools for the analysis of perceptual independence

Description

It fits a GRT-wIND model to data by running grt_wind_fit repeated times, each time with a different value for the starting parameters. It returns the model with a highest log-likelihood from all the runs.

Usage

grt_wind_fit_parallel(cmats, start_params = c(), model = "full",
  rand_pert = 0.3, control = list(), n_reps, n_cores = 0)

Arguments

cmats	List of confusion matrices. Each entry in the list should contain the 4x4 confusion matrix from one individual (see Details).
start_params	An optional vector of parameters to start the optimization algorithm. You can provide either the group parameters or both group and individual parameters. It will be created automatically if not provided (see Details).
model	A string indicating what GRT-wIND model to run. By default is the full model ("full"), but restricted models are also available. "PS(A)" and "PS(B)" are models that assume PS for dimension A and dimension B, respectively. "PI" is a model that assumes perceptual independence. "DS(A)" and "DS(B)" are models that assume DS for dimension A ("PS(A)") and dimension B ("PS(B)"), respectively. "1-VAR" is a model that assumes that all variances are equal to one.
rand_pert	Maximum value of a random perturbation added to the starting parameters. With a value of zero, the algorithm is started exactly at start_params. As the value of rand_pert is increased, the starting parameters become closer to be "truly random."
control	A list of control parameters for the optim function. See optim.
n_reps	Number of times the optimization algorithm should be run. Must be provided.
n_cores	Number of cores to be used. It defaults to all available cores minus one.

Details

rand_pert must be higher than zero. For more details, see grt_wind_fit

Value

An object of class "grt_wind_fit." For more details and examples, see grt_wind_fit

References

Soto, F. A., Musgrave, R., Vucovich, L., & Ashby, F. G. (2015). General recognition theory with individual differences: A new method for examining perceptual and decisional interactions with an application to face perception. Psychonomic Bulletin & Review, 22(1), 88-111.

fsotoc/grtools documentation built on Nov. 15, 2020, 5:14 a.m.