Hands On Monotonic Time Series Forecasting with XGBoost, using Python | by Piero Paialunga | Mar, 2024

That is use XGBoost in a forecasting situation, from principle to apply

A few months in the past, I used to be on a analysis challenge and I had an issue to resolve involving time sequence.

The issue was pretty easy:

“Ranging from this time sequence with t timesteps, predict the following okay values”

For the Machine Studying lovers on the market, that is like writing “Hey World”, as this drawback is extraordinarily well-known to the group with the title “forecasting”.

The Machine Studying group developed many methods that can be utilized to foretell the following values of a timeseries. Some conventional strategies contain algorithms like ARIMA/SARIMA or Fourier Transform analysis, and different extra complicated algorithms are the Convolutional/Recurrent Neural Networks or the tremendous well-known “Transformer” one (the T in ChatGPT stands for transformers).

Whereas the issue of forecasting is a really well-known one, it’s possibly much less uncommon to handle the issue of forecasting with constraints.
Let me clarify what I imply.

You’ve gotten a time sequence with a set of parameters X and the time step t.
The normal time forecasting drawback is the next:

The issue that we face is the next:

So, if we think about that the enter parameter has d dimensions, I would like the perform for dimension 1 (for instance) to be monotonic. So how can we cope with that? How can we forecast a “monotonic” time sequence? The best way that we’re going to describe on this drawback is the XGBoost.

The construction of this blogpost is the next:

1. In regards to the XGBoost: In a number of traces we are going to describe what the XGBoost is about, what’s the elementary thought, and what are the professionals and cons.