I am trying to use a neural network algorithm to solve a statistical problem. When I call the function Nnet (software R) with the following parameters :
rn=nnet(resignation~., data=T[,-1], entropy=T, size=5, decay=1, MaxNWts=3000, maxit=1000)
I have the error below, and I don't succeed to understand the meaning of this error and how to solve it:
weights: 2656
Error in nnet.default(x, y, w, ...) : l'objet (list) ne peut être converti automatiquement en un type 'integer'
Error in nnet.default(x, y, w, ...) : the object (list) can not be automatically converted to an 'integer' type
Could you please help me ?
I remain available for any further information Thanks
It appears that the main problem is the fact that you use the name T as your data frame name. As a consequence, the value received by the entropy argument is a list (data frames can be seen as lists in R) instead of a boolean (or integer). To make it work, run this instead:
rn=nnet(Species~., data=T[,-1], entropy=TRUE, size=5, decay=1, MaxNWts=3000, maxit=1000)
N.B: I would suggest you not to use T as a variable name.