decisiontree.hs

Simple categorical decision tree for Haskell

Installation

The repository provides the necessary configuration for a quick cabal install:

$ git clone https://github.com/wbadart/decisiontree.hs \
    && cabal install decisiontree.hs

Usage

The package exposes the modules DecisionTree and Entropy. The provides the structural algorithms for filtering training instances down the branches of the tree while the latter one, as the name suggests, provides the entropy calculations for branching criteria.

For instance, to run an ID3-like classification on an instance, use:

import DecisionTree (classify)
import Entropy (informationGain)

classification =
    classify trainingInstances informationGain ["red", "hot", "chili"]
        where trainingInstances = someDataYouHave

The package also provides a sample driver program via Main.hs, which compiles to the binary interactivedt (which is currently hard-coded to information gain) which can be used as follows:

$ interactivedt mydata.csv
instance> sushi,CA,cloudy
"positive"
instance> burritos,IN,rainy
...

License

Distributed under the MIT license. See LICENSE and opensource.org for details.