Tuesday Tip #14: Ensemble your ML models for better performance 🎯

Hi Reader,

Last week, I had an hour-long chat with my friend Ken Jee for his podcast. Starting tomorrow, you can watch the conversation on the Ken’s Nearest Neighbors YouTube channel!

👉 Tip #14: How to ensemble your ML models

When working on a Machine Learning problem, it’s always a good idea to try different types of models to see which one performs best.

However, you can also use a process called “ensembling” to combine multiple models. The goal is to produce a combined model, known as an ensemble, that performs better than any of the individual models.

The process for ensembling is simple:

  • For a regression problem, you calculate the average of the predictions made by the individual regressors and use that as your prediction.
  • For a classification problem, you can either average the predicted probabilities output by the classifiers, or you can let the classifiers vote on which class to predict.

The idea behind ensembling is that if you have a collection of individually imperfect models, the “one-off” errors made by each model are probably not going to be made by the rest of the models. Thus, the errors will be discarded (or at least reduced) when ensembling the models.

Ensembling in scikit-learn

Here’s a simple example in which I ensembled Logistic Regression and Random Forests using scikit-learn’s VotingClassifier:

Notice that the accuracy of the ensemble (0.725) is significantly better than the accuracy of either individual model. (Check out the full code here.)

Advice for ensembling

➡️ Ensembling is useful any time model accuracy (or another evaluation metric) is your highest priority. Keep in mind that the ensemble will be less interpretable than the individual models.

➡️ It’s ideal to include at least 3 models in the ensemble.

➡️ It’s important that all models you include are performing reasonably well on their own.

➡️ It’s best if the included models generate their predictions using different processes, since they will be likely to make different types of errors. (This is what makes Logistic Regression and Random Forests good candidates for ensembling!)

If you enjoyed this week’s tip, please forward it to a friend! Takes only a few seconds, and it really helps me out! 🙌

See you next Tuesday!

- Kevin

P.S. Gym rats vs data scientists

Did someone awesome forward you this email? Sign up here to receive data science tips every week!

Learn Data Science from Data School 📊

Join 25,000+ aspiring Data Scientists and receive Python & Data Science tips every Tuesday!

Read more from Learn Data Science from Data School 📊

Hi Reader, I'm really proud of this week's tip because it covers a topic (data leakage) that took me years to fully understand. 🧠 It's one of those times when I feel like I'm truly contributing to the collective wisdom by distilling complex ideas into an approachable format. 💡 You can read the tip below 👇 or on my blog. 🔗 Link of the week Building an AI Coach to Help Tame My Monkey Mind (Eugene Yan) In this short post, Eugene describes his experiences calling an LLM on the phone for coaching:...

Hi Reader, Last week, I recorded the FINAL 28 LESSONS 🎉 for my upcoming course, Master Machine Learning with scikit-learn. That's why you didn't hear from me last week! 😅 I edited one of those 28 videos and posted it on YouTube. That video is today's tip, which I'll tell you about below! 👉 Tip #45: How to read the scikit-learn documentation In order to become truly proficient with scikit-learn, you need to be able to read the documentation. In this video lesson, I’ll walk you through the five...

Hi Reader, happy Tuesday! My recent tips have been rather lengthy, so I'm going to mix it up with some shorter tips (like today's). Let me know what you think! 💬 🔗 Link of the week A stealth attack came close to compromising the world's computers (The Economist) If you haven't heard about the recent "xz Utils backdoor", it's an absolutely fascinating/terrifying story! In short, a hacker (or team of hackers) spent years gaining the trust of an open-source project by making helpful...