| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
for each test. Also changed how we get the rankings of infected
individuals each day.
|
|
infector is ranked among all potential parents. Cleaned up code for the
predicting victims benchmarking test.
|
|
compare true/recovered. Results look good: ml2 finds close-to-actual
parameters and gets ~60% of the infectors right. I will check how many
recovery gets in the top 2 or 3 infectors.
|
|
|
|
|
|
|
|
penalize less for longer distances
|
|
Also playing around with only allowing most likely parent to try and
infect. Still working to see if some variant produces good results that
allow us to determine seeds/non-seeds.
|
|
|
|
|
|
else (unless they are root victims)
|
|
result where optimization minimizes alpha, delta as much as possible
|
|
impossible :/. We may want to try a model where we just assume
infections — can get parameters for the benchmark test.
|
|
|
|
weight_success
|
|
prob_success/failure
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
