summary.bigKRLS: summary.bigKRLS

In rdrr1990/bigKRLS: Optimized Kernel Regularized Least Squares

Description

Summary function for bigKRLS output.

Usage

## S3 method for class 'bigKRLS'
summary(object, degrees = "Neffective",
  probs = c(0.05, 0.25, 0.5, 0.75, 0.95), digits = 4, labs = NULL,
  ...)

Arguments

object	bigKRLS output.
degrees	"Neffective" (default) or "N". If "Neffective" (default), degrees of freedom for t tests reflects degrees of freedom used to obtain regularization parameter, lambda. Neffective = N - sum(eigenvalues/(eigenvalues + lambda)); see e.g. Hastie et al. (2015, 61-68). 'N' is simply the observed sample size (note this is the default for library(KRLS)). Degrees of freedom for t-tests is either Neffective - P or N - P.
probs	Which quantiles of the marginal effects of each x variable should be displayed?
digits	Number of signficant digits.
labs	Optional vector of x labels.
...	ignore

Value

Returns list with "ttests" (Average Marginal Effect estimates, standard errors, t-values, and p values) and "percentiles" (of the marginal effects).

Examples

# y <- as.matrix(ChickWeight$weight)
# X <- matrix(cbind(ChickWeight$Time, ChickWeight$Diet == 1), ncol = 2)

# out <- bigKRLS(y, X, Ncores=1)
# summary(out)

# s = summary(out, digits = 2, labs = c("Time", "ChickWeightDiet"))

# knitr::kable(s[["ttests"]])     # to format with RNotebook or RMarkdown
# knitr::kable(s[["percentiles"]])

# Calcuate p-values using uncorrected standard errors (not recommended)
# summary(out, degrees = "N")

rdrr1990/bigKRLS documentation built on Aug. 2, 2019, 9:13 a.m.