NEWS

nfidd 1.3.1

removed an unused init_R argument from the renewal model (#652)
clarified the joint nowcasting session (#659)
updated the return link to point to the forecasting concepts session (#656)
renamed the mpox figure and credited its sources on the nowcasting slides (#665)

restructured the course to start with delay distributions, moving the probability and Stan introduction to reference and self-study material (#621)
dropped the unused gen_time_pmf argument from simulate_onsets() (#607)
replaced superseded transmute() with mutate() + select() (#606)
used N for the sample size in lognormal.stan to match gamma.stan (#605)
used the quantile_level column for the QRA weights filter (#603)
added a Stan model index reference page (#612)
enabled site search (#614)
added a real-time Rt exercise and callout to the renewal session (#615)
plotted the forecast reproduction number over the forecast horizon (#639)
added an optional dropdown on baselines to the forecasting concepts session (#641)
added missing ensemble forecasting citations and cited Bracher et al. 2021 for over/underprediction (#648, #604)
marked optional content consistently across the sessions and the timetable (#623, #625)
updated learning outcomes to match course content (#611)
corrected an overstated bias claim for ad-hoc censoring fixes (#634)
highlighted printed Stan models in rendered output (#643)
fixed broken LaTeX in the CRPS formula (#642)

made add_delays() more flexible with configurable delay distributions and hospitalisation probability (#584)
added save_warmup argument to nfidd_sample() (#584)
fixed Stan function extraction regex to handle array[] return types (#584)
fixed forecasting model to generate forecast random walk in generated quantities rather than extending estimated parameters (#585)
allowed zero-day delays in censored delay model (#585)
added forecasting to the joint nowcasting with R model (#585)
improved priors in joint nowcasting models (#585)
added Stan reference documentation pages (#586)
updated session content with improvements from SISMID (#587)

adapted nfidd_cmdstan_model() to work with an include path option, and a model file name argument

fixed a bug in convolution function which affected the earliest part of convoluted time series #475.
renamed target_day to origin_day for clarity #465
added nffid_sample() function to speed up default inference #457
replaced vapply with for loop in convolve_with_delay for clarity #433
streamlined the use of logged and natural R #424
added the summarise_lognormal() function for mean/sd summarises #406

In development version of the package and teaching material for teaching in Bangkok in November 2024.

This included a complete redevelopment of the package where what previously were snippets are now functions.

Initial release of the nfidd package and teaching material for teaching in June 2024 in Stockholm.