Issue #36 - Multivariate Time Series models in Keras

💊 Article of the week

In previous issues, we’ve seen how to train a univariate model using XGBoost. That was great, but imagine you want to use multiple features to predict multiple targets. You can’t do that with XGBoost. That’s why in this week’s article we introduce a way of doing so by using Neural Networks with Keras and TensorFlow. First, we introduce a basic way of training them, followed by a series of more advanced functionalities to give you more control over your model. Check it out here:

If you are interested in playing with the code you can find the notebook at the end of the newsletter 😊

If you want to check the previous two parts of this series, you can find them here:

1. Predicting Forex Time Series data using XGBoost

2. Exogenous Variables for Improving Forex Time Series Forecasting with XGBoost

You can read every story from me (and thousands of other writers on Medium) by signing up here.

---

🎓Learn Feature Engineering in Time Series*

Learn how to improve your Time Series models. Here you’ll discover how to:

• Transform time series into structured tables for ML model forecasting.

• Apply direct and recursive methods for multi-step future predictions.

• Impute missing time series data to maintain dataset completeness.

• Generate features from historical data using lags and windowing techniques.

• Encode categorical data and integrate date-time, trend, and seasonality features.

Enroll today and take the next step in mastering the world of data science!

Use the code MLPILLS at checkout to get a 15% discount on any Train In Data courses.

*Sponsored: by purchasing any of their courses you would also be supporting MLPills.

---

🔥Discover more

What is the difference between Univariate and Multivariate models? Discover it here:

Do you have multiple items with some associated characteristics or properties and you want to find how similar they are? This is really useful for Content-based Recommender Systems. Discover the Cosine Similarity metric here:

💡How did you find this issue?

With your feedback, we can improve MLPills. Click on a link to vote:

• 👍 Amazing!

• 😐 It was OK

• 👎 I didn't like it...

For any feedback please let me know at david@mlpills.dev

Thank you!

🪐Play with the notebook

You can download the notebook here. Unfortunately, this is only for paid subscribers. With the paid subscription you will be supporting my work, allowing me to bring the best possible content to you and the community. I really appreciate your support!