Auction: Internal function to manage multiple experts.

In korydjohnson/rai: Revisiting-Alpha-Investing for Polynomial Regression

Internal function to manage multiple experts.

Description

runAuction is the workhorse of the rai package: it takes an initial expert list and runs the Revisiting Alpha-Investing algorithm to greedily fit (optional) polynomials and interactions to data. The term "auction" is the result of multiple experts bidding to perform the test which determines stepwise ordering. This function is not intended to be called directly, but through rai.

Usage

vif(res, y, X, x, n, p, m, TSS, lmFit)

runAuction(
  experts,
  gWealth,
  theData,
  y,
  alg,
  poly,
  searchType,
  m,
  sigma,
  omega,
  reuse,
  nMaxTest,
  verbose,
  save,
  lmFit,
  baseModel
)

Arguments

res	residuals from current model.
y	the response as a single column matrix.
X	covariates in the current model.
x	covariate being tested for addition into the model.
n	number of observations.
p	number of predictors in the current model.
m	number of observations used in subsampling for variance inflation factor estimate of r.squared.
TSS	total sum of squares; considering current residuals to be the response.
lmFit	The core function that will be used to estimate linear model fits. The default is .lm.fit, but other alternatives are possible. Note that it does not use formula notation as this is costly. Another recommended option is fastLmPure from RcppEigen or related packages.
experts	list of expert objects. Each expert is the output of makeStepwiseExpert or makeScavengerExpert.
gWealth	global wealth object, output of gWealthStep.
theData	covariate matrix.
alg	algorithm can be one of "rai", "raiPlus", or "RH" (Revisiting Holm).
poly	logical. Should the algorithm look for higher-order polynomials?
searchType	A character string specifying the prioritization of higher-order polynomials. One of "breadth" (more base features) or "depth" (higher order).
sigma	type of error estimate used in gWealthStep; one of "ind" or "step".
omega	return from rejecting a test in Alpha-Investing.
reuse	logical. Should repeated tests of the same covariate be considered a test of the same hypothesis? Reusing wealth isn't implemented for RAI or RAIplus (effect is negligible).
nMaxTest	maximum number of tests
verbose	logical. Should auction output be printed?
save	logical. Should the auction results be saved? If TRUE, returns a summary matrix.
baseModel	Features to include as the initial model. When NULL, the base model only includes the intercept. baseModel must be specified as a list of desired features. Each list element is a vector of column names or indices, where vectors of length > 1 specify an interaction term of those features. Please check the transformed data using prepareData in order to determine the correct column names and indexes of your desired model.

Value

A list which includes the following components:

formula	final model formula.
y	response.
X	model matrix from final model.
features	list of interactions included in formula.
summary	included if save=TRUE; matrix where each row contains the summary information of a single test.

korydjohnson/rai documentation built on April 21, 2022, 2:59 p.m.