0

So I'm experimenting with a feed forward network, I'm trying to learn how to use neural networks, and I'm some what confused when it comes to selecting features, hidden layers and neurons.

I have 2 datasets that I'm playing around with. Both datasets have 113 features where dataset 1 has about 630 datapoints and dataset 2 has around 3500 datapoints.

The package that I'm using is caret, mostly because of the abundance of neural network types and settings.

user5232061
  • 33
  • 4
  • this might be useful have a look: http://stats.stackexchange.com/questions/21717/how-to-train-and-validate-a-neural-network-model-in-r – Prayalankar Feb 18 '17 at 11:42
  • Moreover: what are you trying to predict? – Tommaso Guerrini Feb 18 '17 at 11:55
  • I'm trying to predict the trend of a stock, 1 if goes up and 0 if it goes down. The input variables are the search frequency index for a couple of google trend words that I pulled out. – user5232061 Feb 18 '17 at 12:05
  • 1
    Possible duplicate of [How to choose the number of hidden layers and nodes in a feedforward neural network?](https://stats.stackexchange.com/questions/181/how-to-choose-the-number-of-hidden-layers-and-nodes-in-a-feedforward-neural-netw) – Sycorax Aug 17 '18 at 03:39

0 Answers0