diva_grid_search_par: Parallelized DIVA Grid Search

In ghonk/catlearn.suppls: Catlearn Supplemental Functions

Description

Runs a grid search over a set of provided parameters and produces averaged response probabilities

Usage

diva_grid_search_par(param_list, num_inits, input_list, fit_type = NULL,
  crit_fit_vector = NULL, state = NULL, procs = NULL)

Arguments

param_list	List of named parameters to be combined and evaluated for the DIVA model
num_inits	Scalar for number of random initializations to be averaged across for the response probability calculation of each parameter combination.
input_list	List of inputs and labels for the grid search ins Matrix of inputs for selected category structure, R (stimuli) x C (features) labels Vector of labels for selected category structure indexed to the input matrx
fit_type	Character specifying the type of fit that is desired. 'best' for the most accurate fit 'crit' for the closest match to a provided vector of response probabilities
crit_fit_vector	Vector of response probabilities for the 'crit' procedure.
state	List of model parameters. Useful for setting parameters that are not subject to the grid search. generate_state is used if no state is provided. Use ?generate_state to examine defaults.
procs	Scalar for number of processors to use in parallelization. Defaults to detectCores() - 2

Value

List consisting of models, response probabilities and best model result.

ghonk/catlearn.suppls documentation built on May 3, 2019, 5:16 p.m.