Multivariate Forecasting

As you'll have picked up by now if you've been paying attention to this book, the field of time-series has made lots of advances within the last decade. Many extensions and new techniques have popped up for applying machine learning to time-series. In each chapter, we've covered lots of different issues around forecasting, anomaly and drift detection, regression and classification, and approaches including traditional approaches, machine learning with gradient boosting and others, reinforcement learning, online learning, deep learning, and probabilistic models.

In this chapter, we'll put some of this into practice in more depth. We've covered mostly univariate time-series so far, but in this chapter, we'll go through an application of forecasting to energy demand. With ongoing energy or supply crises in different parts of the world, this is a very timely subject. We'll work with a multivariate time-series, and we&apos...