HANDWRITTEN DIGIT RECOGNITION BY MNIST

MNIST dataset:

MNIST is a collection of handwritten digits from 0-9. Image of size 28 X 28

Requirements

• Python 3.5 +
• Scikit-Learn (latest version)
• Numpy (+ mkl for Windows)
• Matplotlib

Requirements II

• Use interactive python notebook in Google Collaboratory or Kaggle.

Introduction

MNIST contains 70,000 images of handwritten digits: 60,000 for training and 10,000 for testing. The images are grayscale, 28x28 pixels, and centered to reduce preprocessing and get started quicker.