Challenge 4 - Explore New Forecasting Algorithms

Content

This challenge is about testing previously unseen time series forecasting algorithms.

The purpose of this challenge is to test your ability to use new models for time series forecasting as soon as they become available.

Requirements

You are required to produce a fully reproducible notebook report explaining all the steps to forecast with Bayesian models & GAMs.

You are free to apply the feature sets you created during Challenge 1.

References:
- Bayes Models
- GAMs
- GAMs method

Data

Challenge dataset contains an hourly time series of the M4 competition.