STATISTICAL MODELLING Part IIC

Practical 1: Introduction to R RDS/Lent 2015

R as a calculator
R can be used as a calculator:
> (9.1^3)*sqrt(14)*exp(-5)/log(4)
Help on any R function can be found by typing a question mark followed by the function, e.g.
> ?exp
You will need to use this help facility extensively (and get used to skim-reading to find the relevant bit!).
Note that R is case-sensitive.
The <- symbol is the usual assignment operator in R, though = can also be used. For instance, we can
assign the value 3 to the variable x, and then perform operations on x. Anything which appears after
the hash symbol # is a comment and need not be typed.
> x <- 3
> round(x^2 + log10(x), 3) # try ?round to see what it does
[1] 9.477
> 37 %/% 3 # try ?’%/%’
[1] 12
> 37 %% 3
[1] 1

Creating vectors
The c function (for ‘concatenate’) combines values into a vector.
> x <- c(3, 6, 4, 2)
> x
[1] 3 6 4 2
> length(x)
[1] 4
There is no such thing as a scalar in R; what one might think of as a scalar is treated as a vector of
length 1. Note that R does not distinguish between row and column vectors unlike Matlab.
You can create a vector y with the same entries using y <- scan(). Enter one component per line and
leave a blank line after the last.
A sequence of equally spaced numbers can be created using the seq function. The rep function provides
different ways of repeating vectors.

Operations on vectors
Operations on vectors in R are performed component by component. For example
> x + x
[1] 6 12 8 4
> x*x
[1] 9 36 16 4
> exp(x)
[1] 20.085537 403.428793 54.598150 7.389056
When operations are performed on vectors of different lengths, the shorter vector is cycled until it is the
same length as the longer vector.

> x <- c(3, 6, 4, 2)

> y <- c(1, 2)
> x + y

1
[1] 4 8 5 4
> x*y
[1] 3 12 4 4
> x^y
[1] 3 36 4 4
> y <- 1:3 # same as y <- c(1, 2, 3)
> x + y
[1] 4 8 7 3
Warning message:
In x + y : longer object length is not a multiple of shorter object length
What are the values of x + 2, 3*x and (2 + x)^3?

Indexing vectors
> x <- c(3, 6, 4, 2)
> x[2] # 2nd component of x
[1] 6
> x[c(1, 3)] # 1st and 3rd components of x
[1] 3 4
> x[-1] # All of x except the 1st component
[1] 6 4 2
> x[-(1:2)] # All of x except the 1st two components
[1] 4 2
> x[1:2] <- c(7.1, 3.4) # We can assign values to components
> x
[1] 7.1 3.4 4.0 2.0

Note that after the final command, x has automatically transformed from a vector of integers to a vector
of floating point numbers (these are a way of representing real numbers on computers, though of course
only to a certain degree of accuracy).
We can also index components of a vector using a TRUE / FALSE (logical) vector.
> index_vec <- c(TRUE, TRUE, FALSE, TRUE)
> x[index_vec]
[1] 7.1 3.4 2.0
Logical vectors can also be created using the binary operator < which performs componentwise compar-
isons.
> x
[1] 7.1 3.4 4.0 2.0
> x > 3.6
[1] TRUE FALSE TRUE FALSE
> x[x > 3.6]
[1] 7.1 4.0

Matrices
We can create a matrix using the matrix function.
> A <- matrix(1:8, 2, 4)
> A
[,1] [,2] [,3] [,4]
[1,] 1 3 5 7
[2,] 2 4 6 8
Can you enter the terms by row instead? Rows and columns of matrices can be extracted in the following
way:

2
> A[1, ]
[1] 1 3 5 7
> A[, 3]
[1] 5 6
Note that the rows and columns thus formed are now vectors. We can check this using the very helpful
str function.
> str(A[1, ])
int [1:4] 1 3 5 7
Here we see that A[1, ] is an integer vector of length 4. To create a 2 by 1 matrix, we use
> A[, 2, drop = FALSE]
[,1]
[1,] 3
[2,] 4
Submatrices can be formed by e.g. A[, 1:3]. The diagonal can be extracted using diag. We can perform
many standard operations on matrices.
> A %*% x # matrix vector multiplication
[,1]
[1,] 51.3
[2,] 67.8
> A*A # componentwise multiplication
[,1] [,2] [,3] [,4]
[1,] 1 9 25 49
[2,] 4 16 36 64
> t(A)
[,1] [,2]
[1,] 1 2
[2,] 3 4
[3,] 5 6
[4,] 7 8
> A %*% t(A) # matrix matrix multiplication
[,1] [,2]
[1,] 84 100
[2,] 100 120
The solve function can be used to invert matrices and solve linear systems.
> solve(A %*% t(A))
[,1] [,2]
[1,] 1.50 -1.25
[2,] -1.25 1.05

A few important functions

> x <- c(3, 6, 4, 2)
> sum(x)
[1] 15
> sum(x > 3) # TRUE is treated as 1 and FALSE, 0
[1] 2
> mean(x)
[1] 3.75
> sort(x)
[1] 2 3 4 6
> sd(x) # standard deviation

3
[1] 1.707825
> mean(A) # mean treats A as a vector
[1] 4.5
> colMeans(A)
[1] 1.5 3.5 5.5 7.5
> rowSums(A)
[1] 16 20

How is the standard deviation being calculated? The function cbind ‘glues’ columns of matrices together.
> cbind(1, A)
[,1] [,2] [,3] [,4] [,5]
[1,] 1 1 3 5 7
[2,] 1 2 4 6 8

Generating random numbers

Independent and identically distributed sequences of random numbers are generated with commands like
rnorm, runif, rchisq etc. (normal, uniform, χ2 ). The corresponding density, cumulative distribution
and quantile functions are, e.g. dnorm, pnorm, qnorm.
> x <- rnorm(1000)
> hist(x, freq = FALSE)
> x_vec <- seq(-3, 3, by = 0.1)
> lines(x_vec, dnorm(x_vec), col = "red") # adds lines to an existing plot

What does the following code do?

> X <- matrix(runif(50*1000, min=-1, max=1), 50, 1000)
> hist(sqrt(50) * colMeans(X) / sd(X), freq = FALSE) # sd treats X as a vector
> lines(x_vec, dnorm(x_vec), col = "red")

Experiment with other distributions and other sample sizes.

Exercises
1. Let Z ∼ N (0, 1). Estimate E(Z|{Z ≥ 1}) and E(Z 6 ).

2. What is the upper 5% point of a χ26 distribution?

3. Use R to solve

3a + 4b − 2c + d = 9
2a − b + 7c − 2d = 13
6a + 2b − c + d = 11
a + 6b − 2c + 5d = 27.