Handwritten Digit Classification on MNIST

I implemented logistic regression from scratch for digits classification on MNIST dataset.

Project Description

I implemented a multinomial classification model for classifying digits in MNIST dataset. In MNIST dataset all images are labelled with the respective digit they represent. Each image is 28X28 pixels, and each feature simply represents one pixel's intensity from 0(White) to 255(Black). For evaluation techniques i used here Confusion Matrix, Precision, Recall, and F1 score.