1
votes

Recently, I started reading about the deep learning. Mainly the weights are pre-trained using unsupervised RBM network and after that, they use neural network networks with many hidden layers to address their task. So my question is, Whether we can use DNN with for 2 class classification problem. Thanks to the people who are going to respond.

2

2 Answers

2
votes

Yes, you can do that with a simple logistic regression on top of your hidden layers (whatever you choose for that, RBMs other autoencoders).

1
votes

Absolutely Yes!

As mentioned by Thomas you can use a Logistic Regression as you output layers. Also, another approach is you can use a Sotmax layer with two classes as you output layers.

Good Luck!