Linear Regression with Multiple Covariates

A workflow for building and testing a linear regression model, complete with R code.

This book consists of analysis and model building using data for wine quality. The goal is to build a model that can predict human preference for wine similar to that seen in the test data. Specifically examining the data for white wine, 1599 observations were collected where people rated wine from 0-10 based on how well they liked it. The data consists of a predictor variable (quality) and several covariates.

A linear model will be built and refined using a regression model that starts out using all available regressors. The final model derived is one that drops four of the original covariates based on the results of testing analysis detailed below.

This book is not meant to be a comprehensive review of linear modeling, regression or the associated statistical methods. It is meant to be an overview of the process intended for audiences that have been introduced to the topic(s) but require assistance with the workflow involved.