Need for gap filling approach within mlr3spatiotempcv for time-series data

[mlr-org]

Need for gap filling approach within mlr3spatiotempcv for time-series data

[issue-label-bot]

Issue-Label Bot is automatically applying the label feature_request to this issue, with a confidence of 0.58. Please mark this comment with 👍 or 👎 to give our bot feedback!

Links: app homepage, dashboard and code for this bot.

[pat-s]

I don't see how the methodology in the links relate to this package.

Filling gaps in time series data is a preprocessing task and hence should be covered as a pipelines (mlr3pipelines). Maybe it fits into {mlr3forecasting} but I've really just taken a coarse look.

[pat-s]

@mohammadreza-sheykhmousa Is there a reason why you almost never reply to my replies? Just wondering. Closing here then.

[mohammadreza-sheykhmousa]

Dear @pat-s I am sorry for the late reply and inconvenience I might caused. No there is no reason to that and actually I am very much fund of you guys at mlr3. Just two past weeks I was/am terribly busy with this gap filling approach for time series data for 20 years Landsat imagery for the whole eu.

• I do agree mlr3pipelines is a better fit. and I'll raise the issue there.
• Technically speaking there are some existing approaches for gap filling data and specially for a huge stack of satellite imagery data as:
  • linear approaches like mean (which is pretty lame approach)
  • 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