Neural Network for binary classification

Dear Community,

I have a binary classification problem, with approximately 16 features.

I have been using a logistic regression classifier in MATLAB, and this is performing OK.

I want to try a Neural Network to see if I can improve the classification. Importantly, I want to be able to interrogate the output (the classification) from the NN. For example, for a particular input, was the output 0.51 or 0.99?

Any hints on this would be really appreciated.

I apologise if this question as been addressed elsewhere - please point me to the correct thread if this is the case!

Kind regards,

L

Answer 1

[Srivardhan Gadila]

You can refer to the Deep Learning Toolbox — Examples, List of Deep Learning Layers, Create Simple Image Classification Network & activations

Use an imageInputLayer as an inputLayer to input the features to the network and then define rest of the network with convolution2dLayer or fullyConnectedLayer or other layers from List of Deep Learning Layers followed by softmaxLayer & classificationLayer at the end of the network architecture.

Then use activations function to check the output of softmaxLayer for a given input.