johnnay / datafsm Goto Github PK

original: to_drop <- colnames(data)[-which(names(data) %in% c("period", "outcome"))[nzvs]]

: sometimes NA values are occured(ex) nzvs: 2,3 )

I think this line is right like below.

to_drop <- colnames(data)[-which(names(data) %in% c("period", "outcome"))][nzvs]

I wonder it is right.

Thank you!

http://cran.r-project.org/web/packages/diagram/index.html

I am trying to use your package, starting with the example you gave: fake data example.
but i keep having this error with : res<-evolve_model(cdata)

Error in fitness(Pop[i, ], ...) : unused argument (maxfitness = 1)

any idea why this happens?
thanks.

Because the optimization problem is multi-modal and non-convex and we are using a heuristic optimization procedure, runs can generate different models depending on initialization. Therefore, do random restarts and return best, or return all for the user to select from based on interpretability.