Multiple Regression Analysis

Introduction

The objective of this report is to build an appropriate multiple regression model to predict oil usage of a house, based on a several relevant potential predictors. These potential predictors that can be used in a linear regression model for predicting oil usage are

• Degree Days
• Home Index
• Number People

The main purpose of this analysis is finding out which among these variables should be included in what could be considered the best model. In order to select the best model there are several criteria that will be used, including parsimony, percentage of explained variation in the response variable, size of the standard error.

One way of hand picking a model would be to run a full model, with all the predictors and assess which predictors are individually significant, and only keep those in the final model. Such approach may exhibit some shortcomings (as it does not take into account other elements such as the possibility of the existence of multicollinearity and heteroskedasticity). One way of selecting predictors in automated way is the use of a Stepwise Regression, which will be used in a further section of this report.

Price: $24.64

Solution: The downloadable solution consists of 15 pages, 964 words and 7 charts.
Deliverable: Word Document