##README

This README describes how to compile and run the program as requested in ML Practical Test.pdf:

    dectree input-file output-attribute-index

where the output should be a set of rules defining the built decision tree in the following format:

    Attr1 = Val1 & Attr2 = Val2 ... => AttrN = ValN

Download

Click here to download the entire project as a tarball or a zip file.

Prerequisites

None, basically:

• A C++ compiler supporting the -std=c++11 flag.
• A shell (sh, bash, dash, etc.) to run the compiled program from the command line.

Compile

    cd src && make clean all

Run

    ./dtree ../data/adult.csv 0

Run using shortcuts

    make adult0     Shortcut for: ./dtree ../data/adult.csv 0
    make adult1
    make adult2
    ...

Using a smaller dataset:

    make small0     Shortcut for: ./dtree ../data/adult14.csv 0
    make small1
    make small2
    ...

Other datasets:

    make party0
    make party1
    ...

    make weather0
    make weather1
    ...

Resources

• The Wikipedia article about ID3
• This tutorial explains very well the concepts about about Decision Trees, ID3 , using figures, tables, pseudo-code, and Python code snippets.
• Decision Tree Classifier (C++)
• This Python github project prints out a tree as a flow chart, which makes the whole idea of Decision Trees very easy to grasp