Bidders: Making Bidder Objects

In rai: Revisiting-Alpha-Investing for Polynomial Regression

Description

These functions create objects that manage alpha-wealth. There is only one stepwise "bidder" that manages the global wealth (gWealth) but it can have multiple "offspring" when searching for polynomials. The outer rai function creates one gWealthStep object and one stepwise bidder at the beginning. The stepwise bidder makes a local modification to gWealth, though bidAccepted/bidRejected still call gWealth. More stepwise bidders are created as "scavengers" tied to the global wealth. Defaults are not set because these are internal functions called by rai and runAuction and all arguments are required.

Usage

gWealthStep(wealth, alg, r, TSS, p, reuse, rmse, df)

makeStepwiseBidder(gWealth)

Arguments

wealth	starting alpha-wealth.
alg	algorithm can be one of "rai", "raiPlus", or "RH" (Revisiting Holm).
r	RAI rejects tests which increase R^2 by a factor r^s, where s is the epoch.
TSS	total sum of squares of the response.
p	number of covariates (only used when alg == "RH").
reuse	logical. Should repeated tests of the same covariate be considered a test of the same hypothesis?
rmse	initial (or independent) estimate of residual standard error
df	degrees of freedom of rmse.
gWealth	a global wealth object; output of gWealthStep.

Value

A closure containing a list of functions.

rai documentation built on July 2, 2019, 5:02 p.m.