Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[ENH] N-Hits multivariate forecasting #1729

Open
moghadas76 opened this issue Dec 17, 2024 · 3 comments
Open

[ENH] N-Hits multivariate forecasting #1729

moghadas76 opened this issue Dec 17, 2024 · 3 comments
Labels
enhancement New feature or request feature request New feature or request

Comments

@moghadas76
Copy link

Is your feature request related to a problem? Please describe.
N-Hits multivariate forecasting

Describe the solution you'd like
N-Hits multivariate forecasting

Describe alternatives you've considered
N-Hits multivariate forecasting

Additional context
N-Hits multivariate forecasting

@benHeid benHeid added the enhancement New feature or request label Dec 24, 2024
@benHeid
Copy link
Collaborator

benHeid commented Dec 24, 2024

Hi @moghadas76 can you provide a reference where N-HITS is implemented for multivariate forecasting?

@fkiraly fkiraly changed the title N-Hits multivariate forecasting [ENH] N-Hits multivariate forecasting Jan 5, 2025
@fkiraly fkiraly added the feature request New feature or request label Jan 5, 2025
@Spinachboul
Copy link

Spinachboul commented Feb 18, 2025

@moghadas76 @fkiraly @benHeid
Here is the link for the paper: https://paperswithcode.com/paper/n-hits-neural-hierarchical-interpolation-for
The paper talks about the following points:

  • Introducing N-HiTS, a forecasting model designed to handle long-term predictions more accurately

This is required for the following reasons;

  1. Predictions fluctuate too much
  2. Generally the computational costs are high

How does this help?

  1. Hierarchial Interpolatin
    Helps the model refine predictions in a structured way, focusing on different patterns (high frequency changes vs slow trends)
  2. Multi-Rate Data Sampling
    Allows the model to process data more efficiently by looking at different timescales separately

@Spinachboul
Copy link

Would like to work in this issue if possible, also there is code reference here for N-Hits:
https://github.com/Nixtla/neuralforecast/blob/3f4a837a0f61e65d831d4119769d9da33caa020f/nbs/models.nhits.ipynb#L4

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request feature request New feature or request
Projects
None yet
Development

No branches or pull requests

4 participants