Skip to contents

Select Optimal Number of Synergies (r)

Source: R/r_selection_mixture.R
select_r.Rd

Main entry point for selecting the optimal number of synergies (latent factors) in mixture models. This unified interface dispatches to specialized selection methods based on the specified parameters.

Usage

select_r(
  list_of_data,
  model_type = c("MFA", "MPCA"),
  method = c("vaf", "vss", "smoothness", "cv", "multicriteria"),
  scope = c("global", "clusterwise", "hybrid"),
  K = NULL,
  rvec = 1:6,
  ...
)

Arguments

list_of_data

A list of N matrices, each of dimension (T_i x M)

model_type

Character, either "MFA" or "MPCA"

method

Character, selection method. One of:

  • "vaf" - Variance Accounted For threshold

  • "vss" - Very Simple Structure criterion

  • "smoothness" - Smoothness-based selection

  • "cv" - Cross-validation

  • "multicriteria" - Combined multiple criteria

scope

Character, scope of selection. One of:

  • "global" - Estimate r from pooled data

  • "clusterwise" - Estimate r separately for each cluster

  • "hybrid" - Combine global and clusterwise estimates

K

Integer, number of clusters (required for clusterwise/hybrid scope)

rvec

Integer vector, candidate r values to evaluate. Default: 1:6

...

Additional arguments passed to underlying selection functions

Value

A list containing:

best_r

Selected optimal r value

summary

Data frame with evaluation results for all candidate r values

best_model

Fitted model with optimal r (if refit_final = TRUE)

details

Additional details from the selection method

Details

This function provides a simplified interface to the various r selection methods available in the package. For more fine-grained control, use the specialized functions directly (e.g., select_r_by_vaf_mfa, select_r_mixture).

Deprecated Functions

The following functions are deprecated and will call select_r() with appropriate parameters:

  • select_r_by_vaf_mfa - Use select_r(method="vaf", model_type="MFA")

  • select_r_by_vaf_mpca - Use select_r(method="vaf", model_type="MPCA")

  • select_r_by_vss_mfa - Use select_r(method="vss", model_type="MFA")

  • select_r_by_vss_mpca - Use select_r(method="vss", model_type="MPCA")

See also

select_r_mixture, select_r_by_vaf_mfa, select_r_by_vss_mfa, select_r_by_smoothness_mfa

Examples

if (FALSE) { # \dontrun{
# Simple VAF-based selection for MFA
result <- select_r(list_of_data, model_type = "MFA", method = "vaf",
                   scope = "global", K = 2, rvec = 1:5)

# Multicriteria selection with clusterwise scope
result <- select_r(list_of_data, model_type = "MFA", method = "multicriteria",
                   scope = "clusterwise", K = 3, rvec = 2:6)

# Cross-validation for MPCA
result <- select_r(list_of_data, model_type = "MPCA", method = "cv",
                   scope = "global", rvec = 1:4)
} # }