Image_Recognition

This is nothing but a "Hello World" Program in the Machine Learning World.

I'm creating v1 using only a simple "Logistic Regression and One vs All Classification".I'm expecting accuracy of ~90% or less.

The code name of this project is "IMGRG".

The dataset can be found at IMGRG.

You can find various details in the log.txt file.

Update 1.1

Just Updated the comments in various files to make them more accurate and informative.

Update 1.2

The documentation has been improved.

Python files have become more responsive and informative to the console, with displaying the digit number while creating its data.

The model size has increased and now accomodate all the digits and given digits, except {0}. This would be provided in a later update.

The model has become less accurate the accuracy dropped from 98.72% to 94.82%, owing to the increased size of training data.

The model is now saved as 100 iterations as default. This was the demand of the increasing size of data and decreasing accuracy rate.

The IMGRG data files will not be uploaded as they are very big in size and Github can't support them.