-
For the API design proposal, see [this wiki entry](https://github.com/alan-turing-institute/sktime/wiki/Forecasting-API-proposal).
## Forecasters
### Atomic
- [x] NaiveForecaster (strategies={"l…
-
### Description
To my understanding, currently the implementation of horizon weights is by adding a fixed horizon_weight list when initializing the loss function.
```python
horizon_weight = np.h…
-
# Modeltime Ecosystem Roadmap
The __`modeltime` project roadmap__ tracks the overall development of the Modeltime Ecosystem of forecasting packages. Modeltime is a cutting-edge ecosystem for forec…
-
- PyTorch-Forecasting version: 0.9.0
- PyTorch version: 1.7.1
- Python version: 3.6.13
- Operating System: Amazon Sagemaker Juypternotebook
### Expected behavior
I'm trying to apply a weight to…
-
The following typos are present in the `sktime` repository as of commit 4be3b36. Both of these commands are executed from repository root in a fresh shallow clone.
As identified by `typos --lo…
-
Hi @Shrad-Shukla1 and @Yoon-Eric
(edited on April 30)
Okay, ```(xr_da * aw_factor).mean(dim=[lon_name, lat_name])``` calculates the areal-weighted average per grid cell. see below example.
h…
-
We should collect "simplest non-trivial" examples for the `GraphPipeline` use case - at least one example per @benheid, @miraep8, @fkiraly.
Requirement is full spec with data, intended functionalit…
-
Need to discuss what all we want to add for version 0.3.0
-
**Describe the bug**
In real life, when forecasting a large number of time series, we are faced with very different time series in length. When I made a wide format dataset in which each time seri…
-
An idea for a new wrapper model when we have weights available in the estimation.
Use IRLS as in RLM on top of current MLE models.
e.g. a simple implementation would use an existing model and model c…