Lecture 1: Paper Overview

161.221 Applied Linear Models

Download the R markdown file for this lecture.

This lecture provides an overview of 161.221.

This paper is all about linear models.

• These models seek to describe the variation of one variable in terms of one or more others, so far as this is possible.

• Linear models are a vital tool in the application of statistics.

Learning Outcomes

A student who successfully completes this paper will be able to:

• develop appropriate linear models for data analysis;

• critically assess whether a linear model adequately describes how one or more explanatory variables affect a response variable;

• make inferences about the model parameters, and interpret these in context;

• create and understand analysis of variance tables, and use them to test hypotheses about model parameters;

• compare nested models and select a subset of explanatory variables that explain variation in a response;

• use suitable statistical software to apply linear models.

N.B. the “suitable statistical software” is R. Even if you have used R before, you may wish to review some introductory material to refresh yourself. If you have not used R before, then there is some extra work to do, but we won’t use this software in the first three lectures so there is still time to allow you a decent chance to catch up. Look for links on Stream, including how to get R and RStudio set up for this semester.

Some of the New Stuff You’ll Learn

Polynomial Regression

Model Diagnostic Plots

Response Surfaces

Regression for Grouped Data

Ozone vs temperature for each of five months (May to September)

Computer Practicals

In practice, statistical methods are implemented on computers.

Doing statistical analyses helps you to understand underlying theory.

Computer practicals are an essential element of 161.221.

Practical sessions will be held each week during semester.

Course Synopsis

1. Introduction

2. Simple Linear Regression

3. Multiple Linear Regression

4. Polynomial regression

5. Linear Models with Factors

6. Time Series and Forecasting using Linear Models

161.221 Stream site

Check the Stream site for 161.221 on a regular basis for:

• Announcements and news. These will automatically be emailed to you so do make sure the correct email address is on your Massey profile.

• All study material. There is no set text to purchase.

• Assignments and other assessment exercises, including the submission portal

• Data files for download.

• a record of interactions with staff and classmates

Data download into R

Datasets are also hosted at: https://r-resources.massey.ac.nz/data/161221/

You will see how they can be read directly into R from that location in our practical sessions.

Assessment for 161.221

You will use R for all assessment exercises.

• Assignment 1: You will write this assignment using R markdown.

• Test 1: (Short test) You can choose what software to write your answers, but you will probably need to use R to complete some small tasks.

• Assignment 2: You will write this assignment using R markdown.

• Test 2: (longer test) You will write your answers using R markdown.

R markdown???

If you have not used R markdown in any other courses, then you will need to gain some skill using it. These lectures were written using R markdown. The practical sessions will be done using R markdown template files we provide. You will soon find that R markdown is a huge time saver, especially for anyone who isn’t perfect (probably all of us).

Any Questions About the Course Structure or Direction?

Ask now, on Stream, or by e-mail later… but always ask.