Reputation: 11
rnn()
function R has no return statement. It generates synapses for input, hidden and output layer. How to use these for prediction with a test sample of a time series data?
Upvotes: 1
Views: 877
Reputation: 83
There was just an update of the rnn
package, with the version 0.5.0, it can generalize outside of the toy example of binary addition.
You must use the trainr
function to train the model and the predictr
to predict the values on your data.
So far it only support synchronized many to many learning, understand that each new time point input will produce an output.
Upvotes: 1