Handwritten Digit Image Classifier is a machine learning classifier model for classification of the handwritten digit images.
The reason behind implementaion of this project was to get a better insight into the neural network model of machine learning.
This model is applied using a neural network with input for the model as the pixels of the data and output of the model as the category of the image i.e. the value of the digit.
Implementation of this neural network is done with Octave with algorithms from scratch to get a better understanding of the mechanism of the neural network.
Also, various curves like validation curve, learning curve, degree curve were used for analysis of the behaviour of the model and tuning the parameter of the model.
Tuning the parameter resulted in very good accuracy with the test data.
So, This repository has Octave code with the dataset for this project and excuting the main programme the result will be the accuracy of the model with various learning and validation curve.