README.md

In TanguyPerennec/Autostats: Automated medical science statistics

Autostats

Overview

The goal of autostats is to help you perform basics medical statistics from a dataframe, such as :

1. Table 1 : percentage, median, mean and statistical test between groups (khi2, Fisher, student)
2. Logistic regression : with automated multivariate variable selection

Installation

# The package is only available from github :
# install.packages("devtools")
devtools::install_github("TanguyPerennec/Autostats")

Getting started

library(autostats)

Autostats functions are :

• “Table 1” which render an usual table 1 with basics descriptive statistics (median, mean and proportions) of 2 groups (define by a binary variable called “y”) from a dataframe (called “DF”). Use Table1(DF,y)

• “Logistic regression” which render a matrix with the results (odds ratios, confidence intervals and p-values) of univariate logistic regression of all the variables provided from the dataframe, and the results of a multivariate logistic regression with automated variable selection. Use reglog(DF,y) to have the matrix results or just multivariate_selection(DF,y) to have the selection with details of the selection method. You can provide a different method (“backward”,“forward”,“stepwise”,“augmented”) and selection criteria (“deviance”, “AIC”, “BIC”).

Related work

Autostats will be used in an simulation project of multivariate logistic regression selection so as to provide a better understanding of multivariate selection process.

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