Open AugustHao opened 10 months ago
in dev update: https://github.com/AugustHao/GPreff/commit/87c509d4a284be40b2bf9c9b311efbba2a3d5e86
DOW strength proportional to weekly mean infection count implemented, however it is computationally significantly slower than constant DOW estimation, suspect there might be some way to do this faster.
For the moment 3 DOW options are available: do nothing, estimate a constant DOW, estimate DOW proportional to infection count.
DOW implementation also no longer uses a Dirichlet prior, so init helper works again
working on improving computation speed
make strength of DOW a linear function of total infection count without introducing identifiability issue - will do in a new dev branch