In asfarlathif/RefineMod: Refine Linear Regression Models

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

RefineMod

version <- as.vector(read.dcf('DESCRIPTION')[, 'Version'])                                                                 
version <- gsub('-', '.', version)

The goal of RefineMod is to provide functions to refine and optimize linear regression models. Models can be refined to only those predictors that are statistically significant in explaining the response variable. Linear regression models can also be compared on the basis of their performance like RMSE, R2 and MAE. Package website RefineMod

Package Development

This package was developed as an assignment for STAT545B at UBC using devtools and usethis packages.

The codes and pipeline used to create this package is as follows

usethis::create_package(path = "RefineMod") #To initiate the package project locally

usethis::use_git() #To create a git repository of the package locally

This local repo is then linked to an empty github repository created with the same name RefineMod

git remote add origin https://github.com/asfarlathif/RefineMod.git
git branch -M main
git push -u origin main

usethis::use_r("function name") #To create script files of the functions in this package

#All the functions were documented using the roxygen skeleton

devtools::document() #To record the documentation files and update NAMESPACE

usethis::use_readme_rmd() #README file initiation

usethis::use_mit_license() #LICENSE file

usethis::use_code_of_conduct() #CODE OF CONDUCT file

usethis::use_testthat() #To Create tests to include test scripts for the functions
usethis::use_test("function name") #intialize Funstion specific test scripts

usethis::use_package("package name") #To include package dependencies in the DESCRIPTION file

usethis::use_vignette("vignette name") #Initialize vignette RMD file

usethis::use_news_md() #Adding Changelogs

#PACKAGE DIAGNOSIS

devtools::test() #Run all testthat files
devtools::check() #Head to Toe evaluation of the package

Installation

The development version of RefineMod can be installed from GitHub with:

devtools::install_github("asfarlathif/RefineMod")

Example

Refining a linear regression model with only its significant predictors

cancer_sample data set from datateachr package

radius_mean as response variable and all (except diagnosis) as input predictors

library(RefineMod)
library(datateachr)

mod <- lm(radius_mean ~ ., cancer_sample[,-2]) #lm() call without any predictor selection

summary(mod)

Above is the model built using all the input predictors as the independent variables while building the model. Many of these variables don’t show any statistical significance (in terms of their p-value) to be included in the model.

sig_mod <- lm_significant(cancer_sample[,-2], res = "radius_mean") #model with optimized predictors

summary(sig_mod)

perimeter_mean, compactness_mean, radius_worst, area_worst, concavity_mean, perimeter_worst and compactness_worst are the predictors that were found to be statistically significant while building a model for the response variable radius_mean.

Comparing Model Performance between one or more lm models

train <- mtcars[1:20,]
test <- mtcars[21:30,]

mod1 <- lm(mpg~wt, train)
mod2 <- lm(mpg~cyl, train)
mod3 <- lm(mpg~wt+cyl, train)
mod4 <- lm(mpg~carb, train)

comp_mods(mod1, mod2, mod3, mod4, newdata = test)

Code of Conduct

Please note that the RefineMod project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

asfarlathif/RefineMod documentation built on Dec. 31, 2020, 7:51 p.m.