Martijn Schuemie

@ablack3 : have you had a change to work on this, as discussed? https://github.com/OHDSI/DatabaseConnector/discussions/275

Yeah, I don't like adding the `dbms` parameter, because then any code written using `in_database_schema()` would need to provide that. It seems to me the problem isn't so much that...

Please note that because you have this PR open, any push to the develop branch will trigger Github Actions twice, once for the push to develop, and once for the...

Yes, we're using cross-validation, and in fact there's a difference of 3.469447e-16 in the optimal hyperparameter! But as far as R is concerned, the fitted coefficients are identical (using the...

(BTW, rounding the PS scores sounds like an excellent idea)

No, there's no required time-frame. It would just be good to have this eventually.

Did a first assessment with a 'medium-sized' set (24k covariates, 450k rows), which was not a big success: With KKTSwindle: 8.5 hours Without KKT Swindle: 16 minutes Will try with...

(Sorry for responding in someone else's repo) Seems reasonable, but note that the sin (or cos) is only applied to part1, not part0 (see [here](https://github.com/ojus1/Time2Vec-PyTorch/blob/master/periodic_activations.py#L12) or equation (1) in the...

Hi @pavgra ! There is currently only an [analysis for measurement values](https://github.com/OHDSI/FeatureExtraction/blob/master/inst/csv/PrespecAnalyses.csv#L69), so based on `value_as_number` , not `value_as_concept`. I agree it would be good to also have covariates based...

That is an interesting idea! Because we know the universe of IDs (standard concept IDs in the vocab) we can even compute how often collisions occur beforehand. I would like...