This is my own implementation of a feedforward neural network. It's currently under construction, but I plan to have it support the following:
- A network with one hidden layer.
- Setting the default activation functions. (Default will be the inverse logistic function.)
- Threshold tuning, in the case that the neural network is performing a binary classification problem.
- Neural Networks - A Systematic Introduction (Raul Rojas)
<li> <a href="http://stats.stackexchange.com/questions/181/how-to-choose-the-number-of-hidden-layers-and-nodes-in-a-feedforward-neural-netw">
Cross Validated - How to choose the number of hidden layers and nodes in a feedforward neural network?</a></li>
<li> <a href="http://neuralnetworksanddeeplearning.com/chap2.html">Back Propogation by Michael Nielsen</a></li>
Run the following script to create hard links between the neural network files and the demo scripts located in /demo:
./link.sh
This will also add these links to the ./demo/.gitignore, so they will not be duplicated under version control.
To remove the ./demo/.gitignore and the linked files, run the following:
./link.sh reset