hmelberg / decisionsupport Goto Github PK

another model and veresion, compared to current (for markov)
(or just another version)
for decisiontree: another start node in same tree, or a different tree 8model and version, or just versjon)

also compare against "no treatment" (just a parameter change from the model

and "standard treatment"

have to make all of this easy

use dropdown?

in many senses

• multiple people have access to same model and can change it (one at a time) (share access)
• multiple people can work on same model at the same time (concurrency, cooperate)
• other people can get the model and then copy it and work on in without changing the source model you have (fork)

or the feature may just be deleted ... sort models according to date anyway ... a little redundant

make a stochastic model determnistic by replacing all distribution with its mean values
make a deterministic model stochastic by introducing beta distributions for all p and gamme for all costs and benefits (with standrd error about 20%)

ask user about nature of some columns when importing

parameters
name column, distribution column, mean column, se columns, parameters for dist column etc

make shift distance depen on zoom level, not a constant (80?)

if has done one psa, no need to do it again

• when comparing against anotehrmodel
• when using psa to estimate value of information
• when constructing acceptability curves

decision trees with markov end points