Skip to contents

Load a Pre-defined MVPA Model

load_model.Rd

Retrieves a model specification from the pre-defined set of MVPA models.

Usage

load_model(name)

Arguments

name

Character string specifying the model to load. Must be a pre-defined MVPA model name:

corclass

Correlation-based classifier with template matching

sda_notune

Simple Shrinkage Discriminant Analysis without tuning

sda_boot

SDA with bootstrap resampling

glmnet_opt

Elastic net with EPSGO parameter optimization

sparse_sda

SDA with sparsity constraints

sda_ranking

SDA with automatic feature ranking

mgsda

Multi-Group Sparse Discriminant Analysis

lda_thomaz

Modified LDA for high-dimensional data

hdrda

High-Dimensional Regularized Discriminant Analysis

Value

A list containing the model specification with the following components:

type

Model type: "Classification" or "Regression"

library

Required R package(s) for the model

label

Human-readable model name

parameters

Data frame describing tunable parameters

grid

Function to generate parameter tuning grid

fit

Function to fit the model

predict

Function to generate predictions

prob

Function to generate class probabilities (classification only)

See also

MVPAModels for the complete list of available custom MVPA models

mvpa_model for using these models in MVPA analyses

Examples

# Load custom MVPA model
model <- load_model("sda_notune")

# Load correlation classifier with parameter tuning options
corr_model <- load_model("corclass")
print(corr_model$parameters)  # View tunable parameters
#>   parameters     class                                           label
#> 1     method character correlation type: pearson, spearman, or kendall
#> 2     robust   logical                                   mean or huber