[TOC]
This repository contains two assignments from the course "Fundamentals of Machine Learning Engineering" at Beihang University (BUAA). These assignments utilize Support Vector Machines (SVM) and Convolutional Neural Networks (CNN) to train on the MNIST dataset. The SVM assignment achieved an accuracy of 97.79%, while the CNN assignment achieved an accuracy of 99.55%.
The MNIST dataset is a widely used benchmark dataset in the field of machine learning. It consists of a large collection of 28x28 grayscale images of handwritten digits ranging from 0 to 9. The dataset is divided into a training set and a test set, allowing us to evaluate the performance of our models.
The repository is organized into two main folders, each corresponding to a specific assignment:
In the "MNIST-Classification-with-SVM" folder, you will find an assignment that focuses on training a Support Vector Machine model for digit classification on the MNIST dataset. The SVM algorithm aims to find an optimal hyperplane that separates the different classes in the dataset.
The "MNIST-Classification-with-CNN" folder contains an assignment that explores training a Convolutional Neural Network model for digit classification on the MNIST dataset. CNNs have shown remarkable success in image recognition tasks due to their ability to capture spatial dependencies in the data.
For specific information, please refer to the respective README documents: README for MNIST Classification with CNN and README for MNIST Classification with SVM.