search

mrc-ide / stochasticity-practical

:sparkles::chart_with_upwards_trend::sparkles: Shiny application for stochasticity practical
https://shiny.dide.imperial.ac.uk/stochasticity-practical/
3 stars 2 forks source link

Example 4 #10

Closed joanna-lewis closed 5 years ago

joanna-lewis commented 6 years ago

See materials in old-version.

joanna-lewis commented 6 years ago

a) Looking for output that looks something like this. Students vary beta and nu. The point is that when beta/nu is small the mean over simulations is below deterministic solution because of fade-out. Even when beta/nu is relatively large, the solutions don't match exactly.

screen shot 2018-10-12 at 16 27 03

screen shot 2018-10-12 at 16 21 27

screen shot 2018-10-12 at 16 28 06

joanna-lewis commented 6 years ago

b) Students vary population size (N), and see how deterministic and stochastic solutions compare in each case.

N=400 screen shot 2018-10-12 at 16 32 09

N = 200 screen shot 2018-10-12 at 16 33 44

N = 100 screen shot 2018-10-12 at 16 34 31

N = 50 screen shot 2018-10-12 at 16 36 09

joanna-lewis commented 6 years ago

c) Students change population size to see where fade-outs start to happen. This time we're looking at all the individual simulations, not just the average across simulations.

N = 100 screen shot 2018-10-12 at 16 41 31

N = 50 screen shot 2018-10-12 at 16 42 08

N = 10 screen shot 2018-10-12 at 16 42 43

Then do a parameter plot of the proportion of simulations that have faded-out by end of simulation, against N. screen shot 2018-10-12 at 16 46 14

joanna-lewis commented 6 years ago

d) Epidemic scenario - start from one infected individual, rather than from steady-state level.

Blue line shows proportion of simulations that have faded-out, increasing with time.

screen shot 2018-10-12 at 17 00 14