Lecture 2 : 20 April 2020
In this lecture, we will introduce the main characters in our story: matrices.
A matrix is a just a table of real numbers, like the following:
Above we have a 4 × 3 matrix or (4,3)-matrix, pronounced “4-by-3 matrix”.
4 × 3 or (4,3) is called the shape of the matrix.
More generally, we can talk about m × n or (m,n)-matrices, where m and n are integers.
We always write the number of rows followed by the number of columns. To help you remember this, think of Remote Control : Rows and Columns.
Operations on matrices
We can add two matrices of the same shape, by adding them entrywise:
We can multiply a matrix by a real number:
This is known as scalar multiplication. The scalar refers to the real number that we are multiplying by.
Finally, we can multiply an m × n matrix with an n × p matrix to produce an m × p matrix:
This is known as matrix multiplication to distinguish it from scalar multiplication.
In summary, we have the following operations on matrices:
- Addition : Add two matrices of the same shape
- Scalar multiplication : Multiply any matrix by a scalar
- Matrix multiplication : Multiply an m × n matrix with an n × p matrix
Matrices in Python
We will now go through some basic matrix examples in Python. Let’s define the following matrix and find its shape:
Note the following:
- We need to import the numpy (Numerical Python) library before using it. We import it and use the short-form “np” to refer to it.
- We use numpy arrays to represent matrices.
- Nested square brackets: We create a matrix by feeding a list of lists into the
np.arrayfunction. Observe the order in which we enter the items.
- Commas: we need commas between entries, and also between rows.
Now let’s carry out the operations on matrices that we saw above.
Addition of matrices is simple: just use
Note that I defined matrix $B$ slightly differently from matrix $A$. Both methods work: $A$ is clearer, but $B$ is more compact. Just make sure to put square brackets and commas in the right places!
A * 2.5 works as well.
Older versions of Python don’t let you use
@ for matrix multiplication. Instead you have to use
You should be able to use
@ if you’ve downloaded the latest version of Python, but it’s good to keep
np.dot in mind as you’ll see it quite often when looking at other people’s code.
Numpy allows you to carry out some other operations with
*. Can you guess what happens with each of the following?
It’s ok to use them, but make sure it’s what you want!
Part 1: Find the error
(The cells cannot be evaluated in the slides) Each of the code cells below contains some errors. Can you modify the code in each cell so that it works?
Feel free to delete or make up some entries if you need.
No errors, but is this the output that you expected?
The code works, but is this the right way to compute AB?
- Download and install Anaconda from https://www.anaconda.com/distribution/
- Follow the instructions here to open Spyder, and familiarize yourself with the Editor, the Console and the Variable Explorer.
Try to define some of the matrices we’ve seen above in Spyder, and play around with some matrix operations. If there are any errors, try to figure out what’s wrong.
Feel free to email me if you have any questions!