R Workshop 2 (Term 2)

Introduction

Hello everyone! Here is a chance to explore some of the code and concepts covered in your QS903 lectures.

Specifically, we cover:

• The curve function
• The %*% operator
• The proportional odds model (for ordered categorical outcomes)
• The multinomial model (for unordered categorical outcomes)

You have also covered the multi-level models. These are cool and exciting! However, we are limited to two hours. So this workshop focuses on the models you may use in your first assignments.

The curve function

Imagine you want to plot a function. One way to do this is to define a function,

myInverseLogit <- function(x){
  exp(x)/(1+exp(x)) # this is the inverse logit function, btw
}

generate some value from the function

x <- seq(from = -5, to = 5, by = 0.1)
y <- myInverseLogit(x)

and then plot a line using those values

plot(x, y, type = 'l', ylab = 'myInverseLogit(x)') # type is line 'l'

The curve function combines generating predictions and plotting a line into one command.

curve(expr = myInverseLogit, from = -5, to = 5, n = 100)

Nice. You may remember using the inverse logit function for the logit model. The invlogit function from the arm package you gives us the same predictions.

require(arm)

## Loading required package: arm

## Loading required package: MASS

## Loading required package: Matrix

## Loading required package: lme4

## 
## arm (Version 1.10-1, built: 2018-4-12)

## Working directory is /Users/jamestripp/Desktop

curve(expr = invlogit, from = -5, to = 5, n = 100)

An interesting feature of curve is that it is easy to overlay a curve onto a base R plot.

x <- seq(from = -5, to = 5, by = 1)
plot(x = x, y = invlogit(x))
curve(expr = invlogit, from = -5, to = 5, n = 100, add = TRUE, col = 'pink') # note add is TRUE