-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
1. Priors Shiny App #57
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @gareth-j ,
Thanks for the work. I’ve reviewed the vignette and the model running process and basic output look good to me. There are just two minor points to be noted.
- The vignette is missing a step for building SPDE model and then specifying priors for spatial range parameter and marginal standard deviation parameter, which is:
spde <- INLA::inla.spde2.pcmatern(
mesh = mesh,
prior.range = c(prior_spatial_range, prior_range_probability),
prior.sigma = c(prior_standard_deviation, prior_sd_probability)
)
The priors are specified by changing the values of “prior_spatial_range, prior_range_probability, prior_standard_deviation and prior_sd_probability” based on users’ knowledge, so there should be slider bars to change their values on the Prior ShinyApp. In addition, slide bars should also be included for the prior values of the temporal correlation parameter. The prior for temporal correlation parameter is specified by changing the values of “prior_alpha”, “pg_alpha” in the code below.
rhoprior <- list(theta = list(prior = "pccor1", param = c(prior_alpha, pg_alpha)))
- The vignette uses the Bristol COVID-19 data by directly filtering the entire COVID-19 data within Bristol. An issue of that is the temporal index does not start from 1 and is not continuous, which is not expected by INLA. So the time index needs to be reorganized into the correct format in the final version.
I’ve been working on space-time MCMC this week. Next week I will focus on the prior tutorial and preparing a Bristol dataset set that contains the expected time index, the response variable, and the selected covariates. The tutorial will explain the data, the priors, then run BHM based on different prior values, and finally compare the model outputs by plotting histograms / density curves.
Xueqing
It'd be nice to be able to compare the results of different priors. |
…sible defaults, added more params to be passed into app
Merging in changes from main
Merging in main
Merging in Xueqing's priors tutorial
Merging in main
… from inputs, spde fixed
Merging in async branch
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you @gareth-j . A minor suggestion on the text for temporal priors help in shiny_priors.R. I was wondering if you could replace the current symbol "a" with the symbol "α"
when referring to temporal autocorrelation parameter in the text, as "α" would align better with "alpha". Apart from this, all look good to me.
Xueqing
OK great, thanks! I've made those changes now. |
This will add a new Shiny app for testing of different priors.
Interface
Plotting
Export
Other
Closes #54
Closes #70
Closes #46
Closes #76