Matthijs Hollanders

icon indicating a website link www.quantecol.com.au icon indicating an email [email protected]

icon company Quantecol icon location Ballina, New South Wales, Australia icon location I am broadly interested in conservation and ecology, with a particular focus on hierarchical ecological models.

Results 22 comments of Matthijs Hollanders

Dynamic indexing with imputed nodes

Thanks Daniel! That did the trick. I'm going to get into nimbleFunctions for real!

Feature request: log matrix product

Thanks for the prompt response and tips, @bob-carpenter, I have made the changes to my code for efficiency! FYI, my HMM is a bit more complex but it's good to...

Feature request: log matrix product

Hey, firstly, yes I did catch the bugs. For my model, I firstly just generate new TPMs for each individual because there are very often individual-by-time varying covariates, such as...

Sunset times are off for Australia

Hi, you're correct. -28.865, 153.565 still gives the wrong results. date lat lon sunrise sunset 1 2023-09-15 -28.865 153.565 2023-09-14 09:46:16 2023-09-14 21:39:09

multiply_log(0, 0) has infinite gradient

Any chance this could be addressed? I think accounting for this is making formulating complex HMMs on the log scale a bit more verbose than it needs to be.

New models in flocker

Hi @jsocolar, sounds good. For the single season model, the key data structures are (1) the number of detections per site, primary occasion, and secondary occasion and (2) the waiting...

New models in flocker

Yes. The idea is that there's an overall occupancy state of site $i \in 1:I$ in season $k \in 1:K$; I'll just write it out for $K=1$ for now. Then,...

New models in flocker

Hi Jacob, I'm having a crack at steps 1--3 as outlined above. Pardon me that I'm a bit slow to understand all of this, as I've never worked on an...

"Error in if (varx == 0) { : missing value where TRUE/FALSE needed" with fit$loo() but not loo::loo(fit$draws("log_lik"))

Hey, thanks for the help. traceback() gives me the following: ``` > fit$loo() Error in if (varx == 0) { : missing value where TRUE/FALSE needed > traceback() 8: posterior::autocovariance(sims[,...

"Error in if (varx == 0) { : missing value where TRUE/FALSE needed" with fit$loo() but not loo::loo(fit$draws("log_lik"))

Hey @jgabry, I don't get the error with `r_eff = FALSE`. Thanks so much!