Nearest Neighbor, Decision Tree, and Perceptron
Practice nearest neighbor, decision tree, and perceptron machine learning classifiers.
This project is intended to show how machine learning classifiers function. By testing the various classifiers and seeing their training vs test accuracies we can learn a lot about each of these classifiers and know in which scenarios each work best.
The project can be run and tested in main.ipynb.
Using the datasets and a max depth of 9, we can achieve a maximum training accuracy of 0.833333 and a maximum test accuracy of 0.6475 using 1200 data points.
Using the datasets and 5 epochs, we can achieve a maximum training accuracy of 0.993333 and a maximum test accuracy of 0.815 using 1200 data points.