Back propagation NN on a linear system

Hello, I have a dataset with four input features and their corresponding regression outputs. I want to train a model using back propagation ANN to predict the output values for new input, please how can I do this. One option is to use linear regression, but that does model my data well enough. Another option is to use fitnet function, but I want to actually work with the code and control the parameters like cost function and activation. Please what code can I use to solve this regression problem using back propagation?

Answer 1

[Yash Sharma]

Hi Stan,

You can create your own custom neural network and can have control over all the training parameters. Take a look at below documentation.

• Define Custom Deep Learning Layers: https://www.mathworks.com/help/deeplearning/ug/define-custom-deep-learning-layers.html

Hope it helps!