Authors: Hien Nguyen, Yilin Li, Lyn Peterson

This repo contains the final project for the Statistical Learning course at Reed College taught by Andrew Bray, Fall 2019.

The MNIST database of handwritten digits is a classic dataset for machine learning learners. The goal of this project is to build classifiers to predict what digit (0-9) is displayed in a given image based on 60,000 training images.

Different modeling methods were explored on the MNIST dataset, including:

• Decision Tree
• Random Forest (with bagging)
• Boosted Trees
• Logistic Regression
• KNN
• SVM

The code for all models can be found under the technical-report folder in the code.Rmd file.

Model results are reported and discussed in the technical-report.pdf file under the technical-report folder.

Our presentation can be viewed here.