apoorvagnihotri / mlpy

Implementations of Random Forests, Decision Trees, Linear Regression, K-Nearest Neighbors and Naive Bayes

Intro

This repository contains a naive implementation of Decision Trees and Random Forests. The variant this repository contains is called Classification and Regression Trees.

Requirements

The following method works on Ubuntu 18.04 LTS. You need to install anaconda and start the cli and type in the following command to replicate the environment on your local machine.

conda create --name ENV_NAME --file Requirements.txt

Replace ENV_NAME with an environment name of your choice.

Usage

• See usage/decision-tree/assignment1.ipynb for a commented usage of the implemented CART algo.
• See usage/random-forest/assignment2.ipynb for a commented usage of the implemented random forest.
• See usage/knn/knn.ipynb for a commented usage of the implemented K-Nearest Neighbour Algorithm for Classification and Regression tasks.
• Folder scripts contains python scripts to download the data used in the jupyter notebooks.
• Look at answer_sheet#.md for the assignment answers.

Credits

The implementation I have here, takes ideas from Josh Gordan's implementation of Classification Decision Trees as a part of online video.