nakul7
commented
8 years ago Dear Diggory,
This makes sense but there is no need to include an additional parameter for the shift (\theta) - because the phase is included in the 5 Fourier coefficients.
Best,
Nakul
On 2016.09.29 11:49, Diggory Hardy wrote:
The idea is to decouple the vector transmission model from the human population during warm-up to allow more fitting cycles.
During warm-up the human population is driven by the input EIR (forced), not the dynamic transmission model, so vector-human interaction is in any case one-way (human infectiousness and availability drives the vector model). Saving and averaging human population parameters over several years will let us run the vector model independently.
Vector emergence is currently fit using the input EIR curve and two parameters: a scale parameter and a time shift parameter. Instead the plan is to use a Fourier series with parameters a0, a1/b1, a2/b2 and shift θ, estimated using a standard downhill fitting algorithm.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/SwissTPH/openmalaria/issues/171, or mute the thread https://github.com/notifications/unsubscribe-auth/AGVYk-DsEV6CueuQJqO9IcJFKhJ5TYVkks5qu4mqgaJpZM4KJyuO.
Nakul Chitnis Department of Epidemiology and Public Health Swiss Tropical and Public Health Institute Socinstrasse 57, Postfach, CH-4002 Basel Switzerland Tel: +41 61 284 82 42 Fax: +41 61 284 81 05 Email: Nakul.Chitnis@unibas.ch
The idea is to decouple the vector transmission model from the human population during warm-up to allow more fitting cycles.
During warm-up the human population is driven by the input EIR (forced), not the dynamic transmission model, so vector-human interaction is in any case one-way (human infectiousness and availability drives the vector model). Saving and averaging human population parameters over several years will let us run the vector model independently.
Vector emergence is currently fit using the input EIR curve and two parameters: a scale parameter and a time shift parameter. Instead the plan is to use a Fourier series with parameters a0, a1/b1, a2/b2 and shift θ, estimated using a standard downhill fitting algorithm.