plsRglm (version 1.7.1)

weighted_significance: Compute cross-validation weighted significance scores

Description

This function combines the empirical distribution of selected component counts with a matrix of bootstrap significance indicators. Rows in matind are matched to component counts using names such as "YT1", "YT2", and so on.

Usage

weighted_significance(cv_counts, matind)

Value

Named numeric vector containing weighted significance scores for each predictor. When no row of matind matches cv_counts, the function returns a named vector of NA_real_ values.

Arguments

cv_counts

named vector or table of cross-validation component counts.

matind

matrix-like object whose rows are named by model/component labels and whose columns are predictors. Values are typically logical significance indicators returned by confints2signifind and combined with rbind.

Author

Frédéric Bertrand

https://fbertran.github.io/homepage/

See Also

confints2signifind, cvtable, and signpred.

Examples

cv_counts <- c("1" = 87, "2" = 13)
matind <- rbind(
  YT1 = c(X1 = TRUE, X2 = TRUE),
  YT2 = c(X1 = TRUE, X2 = FALSE)
)
weighted_significance(cv_counts, matind)