What happens if I set negative decays for HawkesExpKern

X-DataInitiative / tick

Module for statistical learning, with a particular emphasis on time-dependent modelling

https://x-datainitiative.github.io/tick/

BSD 3-Clause "New" or "Revised" License

491 stars 108 forks source link

What happens if I set negative decays for HawkesExpKern #524

Open HaochenWang1243 opened 1 month ago

HaochenWang1243 commented 1 month ago

Hi. When I pass in a decay matrix that contains negative entries to HawkesExpKern, no error is reported and fit() is also successful. I'm wondering if the results (score, adjacency, baseline, etc.) I got in this case are sensible or corrupted?

Mbompr commented 1 month ago

That really looks frightening, I have no idea if this is valid or not but definitely I didn't have this in mind while implementing the lib. Use it at your own risks!