Comparative Analysis of Machine Learning Models for Handwritten Digit Recognition

Handwritten digit recognition plays a pivotal role in modern computer vision and its applications inoptical character recognition systems. This paper provides an analysis of different machine learning modelsapplied to classify handwritten digits using the MNIST dataset. The dataset consists of features extractedfrom grayscale images of digits (0-9). This study compares and evaluates the performance of various modelsusing metrics such as accuracy and F1-score. Ensemble modelling techniques are also explored to enhanceprediction accuracy and reliability. The findings underline the effectiveness of advanced models likeconvolutional neural networks in digit classification tasks and their potential in practical applications.