boehringer-ingelheim / oncomsm Goto Github PK
View Code? Open in Web Editor NEWBayesian multi-state models for the analysis of oncology trials
Home Page: https://boehringer-ingelheim.github.io/oncomsm/
License: Other
Bayesian multi-state models for the analysis of oncology trials
Home Page: https://boehringer-ingelheim.github.io/oncomsm/
License: Other
Restricting a complete data set to previous time points is a lot easier when working with visit data (just filtering) than with multistate data.
Can we reuse the same .impute() implementation for both?
If not, should we switch to generating visits by default (we can already convert them to multistate data)
We should try to stick to this as close as possible and aim at 0 lintr findings (or use # nolint where necessary)
add missing references and polish code in the getting started and probability of scuccess vignettes.
Currently, sampling under fixed parameters is a bit clunky by reducing the prior variability to close to 0. Would be clearer to allows sampling from fixed parameters.
Time-dependent covariates are a bit tricky to incorporate since they would require a joint longitudinal / multi-state model. Sex, treatment line etc. are fixed at baseline though and can be incorporated relatively simply. Since the model is not hazard-based, covariates would affect logodds of transition probabilities and the location parameter of the transition times.
uniform might not be ideal
oncomsm/inst/stan/srp_model.stan
Line 105 in 9bd2337
The forward sampling is still implemented in R quite naively. We do not want to implement it in stan to keep inference and predictive sampling separate but this should really be implemented in C++ to speed things up a bit.
Lines 65 to 187 in 59dbee9
To work with rcpp we need to create the columns separately and then put them together as data.frame at the very end.
We should gradually increase the unit-test coverage up to 100%. Structural testing is done in vignettes for now, maybe additional pkgdown articles are required as well.
Logodds might be more convenient to implement hierarchical borrowing; for now beta priros might make it more convenient to elicit priors though.
Need to implement own print method to avoid cluttering the console with the stan model.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.