Prior Ridge Lambda Cross Validation — pridge_lambda

Runs leave-one-out cross validation across a grid of lambda and returns the MSEs across the whole grid. This uses the 'foreach' package for parallel execution.

Usage

pridge_lambda_cv(
  design,
  response,
  priors,
  grid,
  plot_mses = TRUE,
  n_cores = NULL
)

Arguments

design: (matrix, or coercable to matrix) 1-hot encoded team lineups, matches on the rows, teams on the columns
response: (vector) response vector corresponding to `design`; typically alliance scores
priors: (vector) length equal to ncol(design) representing a best guess at the coefficient without match data. Typically pre-event EPA.
grid: (vector) all lambda values (regularization parameter) to consider
plot_mses: (boolean) if TRUE, output a plot showing the CV results
n_cores: (int) the number of cores to parallelize over; or NULL to use the max minus 1.