topic Neural Network in SAS Enterprise Miner in SAS Data Science

Neural Network in SAS Enterprise Miner

lokendra_devangan_corecompete_com — Wed, 09 May 2018 14:18:30 GMT

My question is specific to SAS E-Miner. .

1. What is meant by number units in hidden layer? Is it number of neurons in a layer?

2. I am using Neural Network node. How do I know how many layers have been used in the model ?

thanks

Lokendra

Re: Neural Network in Eminor

WendyCzika — Wed, 09 May 2018 14:13:38 GMT

Yes, hidden units and hidden neurons are just different terms for the same thing.

The Neural Network node uses 1 hidden layer. You can use the HP Neural node to include up to 10 hidden layers.

Re: Neural Network in Eminor

lokendra_devangan_corecompete_com — Wed, 09 May 2018 14:16:18 GMT

Thanks.

So, if I get 10 parameter estimates number of neurons is 10, right?

Re: Neural Network in Eminor

WendyCzika — Wed, 09 May 2018 14:19:36 GMT

By default, the Neural Network node uses 3 hidden units/neurons in the hidden layer, but you can change this by clicking on the ellipsis next to the "Network" property and setting the Number of Hidden Units. If there are 10 hidden units, you would have [(# of inputs) * 10 + (# of targets) * 10] parameter estimates (1 estimate or weight for each link from input to hidden nodes, and hidden to output nodes).

Re: Neural Network in Eminor

lokendra_devangan_corecompete_com — Wed, 09 May 2018 14:30:32 GMT

Thanks again. Based on that I should get 6 but I am getting 10. I have only one input. BIAS I understand what are others

N Parameter Estimate Function

1 housing_H11 -1.523604 -0.000000258
2 housing_H12 0.673423 -0.000002911
3 housing_H13 0.472912 -0.000004649
4 BIAS_H11 0.129327 -0.000003474
5 BIAS_H12 -1.482436 -0.000000343
6 BIAS_H13 0.421266 -0.000003785
7 H11_good_badgood 0.713612 0.000000562
8 H12_good_badgood -1.245800 0.000002289
9 H13_good_badgood 1.768447 -0.000001798
10 BIAS_good_badgood -0.803905 -0.000001713

Re: Neural Network in Eminor

WendyCzika — Wed, 09 May 2018 15:01:10 GMT

Sorry, I wasn't including the bias terms in the number of parameter estimates. You would have one of those for each hidden and output node as well.