search

aflaxman / gbd

Generic Disease Modelling System
http://www.globalburden.org
15 stars 10 forks source link

EMP prior maps #67

Closed Farshad-xx closed 14 years ago

Farshad-xx commented 14 years ago

For developing EMP prior maps, seems that DISMOD does not take into account the parameter unit and treats all like with a default unit of 100 (even if unit = 1 has been entered in data).

aflaxman commented 14 years ago

can you add id for the model where you observed this?

aflaxman commented 14 years ago

i think this is fixed in jiaji's latest updates

Farshad-xx commented 14 years ago

I saw yesterday that giving all parameter values with unit one and unit = 1 works better than having different units. So dismod does not handle different units properly now but the safest way is to give all parameter values with unit one in data.

aflaxman commented 14 years ago

can you check this on the new dismod (http://winthrop.ihme.washington.edu:8000); this should handle mixed units properly now, and if it's still broken, I'd like to know.

Farshad-xx commented 14 years ago

Model 2688 uses different units for parameter values and has recognized only the prevalence data but not the other parameters. Model 2687 uses the same data but with unit one before entering to DisMod. All parameters are recognized. So unit still has room for correction.