A Comparative Analysis of Various Machine Learning Algorithms for Handwritten Digit Recognition

One of the most fundamental uses of machine learning is the identification of handwritten digits. There has been a lot of compelling research in this field. Since everyone has different handwriting, it becomes very important for the algorithm to recognize all styles and ways of writing. With the advancement in science and technology, digitization of all forms of documents has become prominent. Since there are still many documents that are yet to be digitized, these applications come in handy. MNIST and SKLearn load_digits datasets are prominent for the recognition process and they have a large set of data points. Many Algorithms are developed over the years that are capable of performing recognition of digits. The accuracy and performance metrics of the methods Support Vector Machine (SVM), Random Forest, Gaussian Naive Bayes, and Convolutional Neural Networks (CNN) are analyzed in this study.