Skip to main content

Course

Intermediate Regression with statsmodels in Python

IntermediateSkill Level

4.8+

Updated 05/2022

Learn to perform linear and logistic regression with multiple explanatory variables.

Start Course for Free

PythonProbability & Statistics4 hr14 videos52 Exercises4,300 XP15,540Statement of Accomplishment

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.

Loved by learners at thousands of companies

Training 2 or more people?

Try DataCamp for Business

Course Description

Linear regression and logistic regression are the two most widely used statistical models and act like master keys, unlocking the secrets hidden in datasets. In this course, you’ll build on the skills you gained in "Introduction to Regression in Python with statsmodels", as you learn about linear and logistic regression with multiple explanatory variables. Through hands-on exercises, you’ll explore the relationships between variables in real-world datasets, Taiwan house prices and customer churn modeling, and more. By the end of this course, you’ll know how to include multiple explanatory variables in a model, discover how interactions between variables affect predictions, and understand how linear and logistic regression work.

Prerequisites

Introduction to Regression with statsmodels in Python

1

Parallel Slopes

Extend your linear regression skills to parallel slopes regression, with one numeric and one categorical explanatory variable. This is the first step towards conquering multiple linear regression.

Parallel slopes linear regression

Fitting a parallel slopes linear regression

Interpreting parallel slopes coefficients

Visualizing each explanatory variable

Visualizing parallel slopes

Predicting parallel slopes

Predicting with a parallel slopes model

Visualizing parallel slopes model predictions

Manually calculating predictions

Assessing model performance

Comparing coefficients of determination

Comparing residual standard error

2

Interactions

Explore the effect of interactions between explanatory variables. Considering interactions allows for more realistic models that can have better predictive power. You'll also deal with Simpson's Paradox: a non-intuitive result that arises when you have multiple explanatory variables.

Models for each category

One model per category

Predicting multiple models

Visualizing multiple models

Assessing model performance

One model with an interaction

Specifying an interaction

Interactions with understandable coeffs

Making predictions with interactions

Predicting with interactions

Manually calculating predictions with interactions

Simpson's Paradox

Modeling eBay auctions

Modeling each auction type

3

Multiple Linear Regression

See how modeling and linear regression make it easy to work with more than two explanatory variables. Once you've mastered fitting linear regression models, you'll get to implement your own linear regression algorithm.

Two numeric explanatory variables

Interactive 3D scatter plot

Visualizing three numeric variables

Modeling two numeric explanatory variables

Visualizing two numeric explanatory variables

Including an interaction

More than two explanatory variables

Visualizing many variables

Different levels of interaction

Predicting again

How linear regression works

The sum of squares

Linear regression algorithm

4

Multiple Logistic Regression

Extend your logistic regression skills to multiple explanatory variables. You’ll also learn about logistic distribution, which underpins this form of regression, before implementing your own logistic regression algorithm.

Multiple logistic regression

Logistic regression with two explanatory variables

Logistic regression prediction

Visualizing multiple explanatory variables

Confusion matrix

The logistic distribution

Cumulative distribution function

Inverse cumulative distribution function

Logistic distribution parameters

How logistic regression works

Likelihood & log-likelihood

Logistic regression algorithm

Congratulations!

Intermediate Regression with statsmodels in Python

Course
Complete

Earn Statement of Accomplishment

Add this credential to your LinkedIn profile, resume, or CV
Share it on social media and in your performance reviewEnroll Now

Don’t just take our word for it

*4.8

from 528 reviews

84%

16%

1%

0%

0%

Sort by

Jorge Luis

2 days ago

Hotatsu

5 days ago

Ammar

last week

h

Camilo

2 weeks ago

Paula

2 weeks ago

Dimitris

2 weeks ago

Jorge Luis

Hotatsu

"h"

Ammar

FAQs

What prior regression knowledge do I need for this course?

You should first complete Introduction to Regression with statsmodels in Python. This course builds directly on those skills by adding multiple explanatory variables and interaction effects.

What real-world datasets are used in the exercises?

You work with Taiwan house prices and a customer churn dataset, among others, to explore relationships between multiple variables using linear and logistic regression.

Does this course cover both linear and logistic regression?

Yes. The first three chapters focus on multiple linear regression, including parallel slopes and interactions. The final chapter extends these concepts to multiple logistic regression.

What is Simpson's Paradox and will I encounter it here?

Simpson's Paradox is a counterintuitive result where a trend reverses when data is grouped differently. You explore it in Chapter 2 while learning about interaction effects between variables.

Will I implement regression algorithms from scratch?

Yes. After mastering model fitting with statsmodels, you implement your own linear regression algorithm in Chapter 3 and your own logistic regression algorithm in Chapter 4.

Join over 19 million learners and start Intermediate Regression with statsmodels in Python today!

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.

Grow your data skills with DataCamp for Mobile

Make progress on the go with our mobile courses and daily 5-minute coding challenges.