reglog: Logistic Regression Function

In TanguyPerennec/Autostats: Automated medical science statistics

Description

reglog is able to perform logistic regression with variable selection and gives as a result a matrix with variable names, odds-ratios, confidence intervals and p-values of univariate and multivariate models.

Usage

reglog(
  DF,
  y,
  explicatives = colnames(DF)[colnames(DF) != y],
  alpha = 0.05,
  dataprep = TRUE,
  verbose = TRUE,
  alpha_max = 0.2,
  round = 3,
  rowstimevariable = 10,
  confirmation = FALSE,
  keep = FALSE,
  exit = "html",
  stability = FALSE,
  equation = FALSE,
  title = TRUE,
  ...
)

Arguments

DF	dataframe, matrix or tibble that contains all explicatives variables and the variable to explain
y	character : name of the variable to explain
explicatives	character vector : variables that should explain y in the logistic regression. Takes all columns but y from the dataframe if kept empty.
alpha	num : significance threeshold used to delete non-significant variables in the multivariate model.
verbose	logical : if TRUE, explainations are displayed in the console while running the function.
alpha_max	num : maximum threeshold used to select the minimum multivariate variables wanted.
round	num : number of digits to display in the final table.
keep	all the variables that should be kept in the multivariate results
exit	specify where do you want to display the results : console (the default), excel (in a results.xlsx file), html (using kable)
method	the method that will be used to select variables in the multivariate model.The default method is the backward elimination. See 'details' section for more informations.

Value

reglog returns a matrix with all OR obtain from univariate model and OR obtain from the multivariate model

References

Bursac, Z., Gauss, C.H., Williams, D.K. et al. Purposeful selection of variables in logistic regression. Source Code Biol Med 3, 17 (2008). https://doi.org/10.1186/1751-0473-3-17

Heinze G, Schemper M. A solution to the problem of separation in logistic regression. Stat Med. 2002;21(16):2409-2419. doi:10.1002/sim.1047

TanguyPerennec/Autostats documentation built on Dec. 13, 2020, 10:43 a.m.