Join us and get access to hundreds of tutorials and a community of expert Pythonistas.

Unlock This Lesson

This lesson is for members only. Join us and get access to hundreds of tutorials and a community of expert Pythonistas.

Unlock This Lesson

Hint: You can adjust the default video playback speed in your account settings.
Hint: You can set the default subtitles language in your account settings.
Sorry! Looks like there’s an issue with video playback 🙁 This might be due to a temporary outage or because of a configuration issue with your browser. Please see our video player troubleshooting guide to resolve the issue.

Multiple Linear Regression: Code

Here’s the data that you’ll need to implement Multiple Linear Regression in Python:

x = [[0, 1], [5, 1], [15, 2], [25, 5], [35, 11], [45, 15], [55, 34], [60, 35]]
y = [4, 5, 20, 14, 32, 22, 38, 43]

00:00 Let’s start again by first importing NumPy and then the LinearRegression class from the sklearn.linear_model module. As before, we’re going to be using some dummy data.

00:14 I’m going to copy and paste some values for the inputs x and the outputs y, and this data will be provided to you underneath the notes that accompany the video lesson.

00:28 So from the data, you see that what we have is eight observations. So eight observations for the responses, and we’ve got eight input observations. Each input observation consists of two data points.

00:44 So we’re using a two-dimensional linear regression model. Let’s take these Python lists and create NumPy arrays out of them. So go ahead and run that and then redefine x to be a NumPy array … and the same thing for y.

01:05 Let’s take a look at the shape of x and the shape of y.

01:14 And so we see this is consistent with some of the notation that we were using in the previous lesson. The input array is a two-dimensional array containing eight rows and two columns. Eight corresponds to the number of observations, and y is a one-dimensional array containing eight components.

01:35 Now let’s use the LinearRegression class and the .fit() method to build the model and compute the coefficients all at once.

01:52 Let’s compute the value.

02:08 This is a good value for , and it tells us that the linear model is a pretty good fit to the actual data. Now, let’s go ahead and print out the coefficients—first the intercept …

02:26 and then the coefficients multiplying the input components.

02:35 So as before, the intercept is one value, one of those coefficients, and then the coefficients multiplying the input components is going to be two because this is a two-dimensional linear model.

02:50 Now let’s compute the estimates for the responses that correspond to the input observations. This time, let’s just call this y_est for y estimate.

03:02 And so these are going to be the values of the model evaluated at the observations. Oh, and this time it wasn’t a t that I forgot. It was a c.

03:16 So let’s compare the estimates and the actual observed values for the responses. So for the first response of four, we estimated a value of 5.7, approximately. For the response actual response of five.

03:37 We estimate with the model a value of about 8.0 and so on. Now let’s use the model to predict responses at new inputs. So go ahead and create an input array.

03:51 Let’s create a five-by-two input array. So you’ll have to use, say, the arange() function just to create some dummy data. So we’ll pass into arange() 10 so that we get the values from zero to nine, and then we want to reshape that so that we get a five-by-two NumPy array. And let’s just pass in the -1 shortcut.

04:16 So just to clarify, when we print this, we’re going to be generating a five-by-two NumPy array or a five-by-two matrix. And then we can predict what the responses are going to be at those inputs.

04:37 And that’s all there is to it. So implementing multiple linear regression is very similar to just simple linear regression. The only difference is making sure that all your dimensions add up.

04:51 In the next couple of lessons, We’ll take a look at how we can implement polynomial regression in scikit-learn.

Become a Member to join the conversation.