leejiaj / backpropagation-algorithm-neural-networks

Implementing the backpropagation algorithm for Neural Networks

Backpropagation-Algorithm-Neural-Networks

Implementing the backpropagation algorithm for Neural Networks

This python program implements the backpropagation algorithm for Neural Networks. There are two steps:

1. Pre-processing the dataset. The two arguments for the program:

• input path of the raw dataset
• output path of the pre-processed dataset

2. Training a Neural Network - Uses the processed dataset to build a neural network. The input parameters to the neural net are:

• input dataset – complete path of the post-processed input dataset
• training percent – percentage of the dataset to be used for training
• maximum_iterations – Maximum number of iterations that the algorithm will run. This parameter is used so that the program terminates in a reasonable time.
• number of hidden layers
• number of neurons in each hidden layer

Pandas is used for reading/pre-processing data.