Robust gap filling approach for time-series data

mohammadreza-sheykhmousa commented 3 years ago

Hi,

Can we have a gap filling approach within mlr3pipelines?

Technically speaking there are some existing approaches for gap filling data and specially for a huge stack of satellite imagery data as in the libraries such as Gdalcube, which includes:
- linear approaches like mean
- repeat last seen observation
- etc. but a robust method for such a challenge is missing. That is why I asked about develop/adopting an approach to address the challenge.

Best, Mo Here are some examples of the problem and solutions for your reference: LINK1 LINK2 LINK3

pfistfl commented 3 years ago

I think this should go into (https://github.com/mlr-org/mlr3raster) @be-marc ?

mohammadreza-sheykhmousa commented 3 years ago

I did :)

mlr-org / mlr3pipelines