Suggested a new section -- Machine learning of spatial data

[Nowosad]

Dear @rsbivand -- I propose adding a new section to the task view "Machine learning of spatial data". It is based on the work we've done for the FOSSGIS 2025 conference -- https://loek-rs.github.io/FOSSGIS2025-examples/. Please review it and let me know how to improve it.

[rsbivand]

@Nowosad Thanks, a good idea! The text could be longer, maybe pointing out that assuming that spatial autocorrelation (whether immanent or driven by arbitrary entitation) is risky when splitting out training and test sets, as information leaks between them, tainting the test sets. I feel that this means that for spatial data this means that not using spatial packages to drive sampling is going to lead to suboptimal inferences. Is there any text in https://rsbivand.github.io/nem24_talk/ you might use?

[Nowosad]

@rsbivand -- I just added a bit of introduction and included one more package. Let me know what you think about it and feel free to edit it.

[rsbivand]

Looks good! I'd go with this now, but if you post again about the addition, maybe solicit views/PRs.

[Nowosad]

Thanks -- I will share it on fosstodon, etc. asking for PRs in the coming weeks.