Mismatch between section subtitles/comments in EpiNow2 howto

[jamesmbaazam]

The guide on how to reconstruct infection dynamics from incidence data on delayed outcomes like deaths will need some alignment in titles and code. In particular,

Example 2 is titled "Reconstruct Ebola infection dynamics in the UK from data on deaths, 1976". The data is however from Zaire. The analysis also uses cases instead of deaths. If going by cases, it'll be a mismatch with the title, so consider broading the main title or using deaths in the guide, although that will require specifying the delay from infection to death.

As an aside, could you explain or note in the "steps in detail" for Example 2 why you used the zero-centered prior (backcalc = backcalc_opts(prior = "none"))?

cc: @adamkucharski

[adamkucharski]

Thanks for flagging. Have created PR #92 to fix the heading. I remember we had some issues with the Ebola reconstruction initially – perhaps backcalc = backcalc_opts(prior = "none") was a suggestion to fix as starting point @sbfnk?

[sbfnk]

prior = "none" centres the Gaussian Process prior on zero which may well not be the best choice - both prior = "reports" (GP prior centred on shifted cases) and prior = "infections" (GP prior on first differences centred on zero) should be easier to fit unless the data is very sparse/noisy.