This project consists of a set of training examples containing Hand Written numeric data. A Neural Network is used to train and algorithm to learn that data and tell us (predict) what number is written should a new number be inputed to test.
This Hand-Written digit recognition finds its application in postal code recognition where thousands of postal codes are scanned and the input digit is then predicted and the postal code is then stored in a database.
The code is written in Octave but will work for Matlab as well.
To run the code simply run ex3_nn.m in Octave or Matlab and let the magic unfold.