Learn learn how to optimize model hyperparameters and even the architecture in a couple of lines of code
In my previous article, we explored the fundamentals of time series forecasting with sktime, learn how to leverage this powerful library for straightforward forecasting tasks. Now, it’s time to take our journey further and dive into the advanced techniques that may enable you to optimize your forecasts and improve their accuracy. On this follow-up, we’ll explore learn how to construct more sophisticated models, tune hyperparameters, and even do model architecture search with sktime.
First, for a simple start, let me show the essential sktime workflow again. This time, we’ll use the Longley dataset, which is a part of sktime (BSD-3 license). It accommodates various US macroeconomic variables from the years 1947 to 1962 and appears like this:
