In R , Why predicted values of a neural network keep changing everytime when we train the network?

Question

Everytime we train the neural network in R (under neuralnet package) for the same data,it predicts different values which aren't the same as predicted earlier ... Any way to solve this issue ???

Answer 1

There is probably some randomization involved in initialization and/or training - you may want to read through the references provided in the documentation.

Try setting the random number generator's seed prior to training, by using set.seed(). (Just don't simply use set.seed(123) ;-).)