MNIST-Digit-Recognition

• A personal project to learn more about Neural Networks and how they operate given the well-known MNIST Handwritten Digit Dataset. Currently, the model uses a Convolutional Neural Network to achieve 98.175% accuracy on Kaggle Testing. I am also currently trying to improve this to 98.5-99% accuracy through optimizations. The CSV output files and corresponding code are used to create accuracy checks on Kaggle.com.

• Link to the corresponding Kaggle dataset/challenge : https://www.kaggle.com/c/digit-recognizer/