chefboost 0.0.7

Lightweight Decision Tree Framework Supporting GBM, Random Forest and Adaboost

chefboost

Chefboost is a lightweight gradient boosting, random forest and adaboost enabled decision tree framework including regular ID3, C4.5, CART, CHAID and regression tree algorithms with categorical features support. It is lightweight, you just need to write a few lines of code to build decision trees with Chefboost.

Installation

The easiest way to install Chefboost framework is to download it from from PyPI.

pip install chefboost

Installation guideline is captured as a video here.

Usage

Basically, you just need to pass the dataset as pandas data frame and tree configurations after importing Chefboost as illustrated below. You just need to put the target label to the right. Besides, chefboost handles both numeric and nominal features and target values in contrast to its alternatives.

from chefboost import Chefboost as chef
import pandas as pd

df = pd.read_csv("dataset/golf.txt")

config = {'algorithm': 'ID3'}
model = chef.fit(df, config)

Outcomes

Built decision trees are stored as python if statements in the tests/outputs/rules directory. A sample of decision rules is demonstrated below.

def findDecision(Outlook, Temperature, Humidity, Wind, Decision):
   if Outlook == 'Rain':
      if Wind == 'Weak':
         return 'Yes'
      elif Wind == 'Strong':
         return 'No'
      else:
         return 'No'
   elif Outlook == 'Sunny':
      if Humidity == 'High':
         return 'No'
      elif Humidity == 'Normal':
         return 'Yes'
      else:
         return 'Yes'
   elif Outlook == 'Overcast':
      return 'Yes'
   else:
      return 'Yes'

Testing for custom instances

Decision rules will be stored in outputs/rules/ folder when you build decision trees. You can run the built decision tree for new instances as illustrated below.

test_instance = ['Sunny', 'Hot', 'High', 'Weak']
model = chef.fit(df, config)
prediction = chef.predict(model, test_instance)

You can consume built decision trees directly as well. In this way, you can restore already built decision trees and skip learning steps, or apply transfer learning. Loaded trees offer you findDecision method to test for new instances.

moduleName = "outputs/rules/rules" #this will load outputs/rules/rules.py
tree = chef.restoreTree(moduleName)
prediction = tree.findDecision(['Sunny', 'Hot', 'High', 'Weak'])

tests/global-unit-test.py will guide you how to build a different decision trees and make predictions.

Model save and restoration

You can save your trained models.

model = chef.fit(df.copy(), config)
chef.save_model(model, "model.pkl")

In this way, you can use the same model later to just make predictions. This skips the training steps. Restoration requires to store .py and .pkl files under outputs/rules.

model = chef.load_model("model.pkl")
prediction = chef.predict(model, ['Sunny',85,85,'Weak'])

Sample configurations

Chefboost supports several decision tree, bagging and boosting algorithms. You just need to pass the configuration to use different algorithms.

Regular Decision Trees ID3 Video, C4.5 Video, CART Video, CHAID Video, Regression Tree Video

config = {'algorithm': 'C4.5'} #ID3, C4.5, CART, CHAID or Regression

Gradient Boosting Video

config = {'enableGBM': True, 'epochs': 7, 'learning_rate': 1}

Random Forest Video

config = {'enableRandomForest': True, 'num_of_trees': 5}

Adaboost Video

config = {'enableAdaboost': True, 'num_of_weak_classifier': 4}

Feature Importance - Video

Decision trees are inherently interpretable and explainable algorithms. Still we can add some extra layers to explain the built models. Herein, feature importance is one of the most common way to make transparent models.

if __name__ == '__main__':
   config = {'algorithm': 'C4.5', 'enableParallelism': True}
   model = chef.fit(df, config)
   fi = chef.feature_importance()
   print(fi)

This returns feature importance values in the pandas data frame format.

feature	final_importance
Wind	0.609868
Humidity	0.265105
Temperature	0.197528
Outlook	-0.072501

Paralellism

Chefboost offers parallelism to speed model building up. Branches of a decision tree will be created in parallel in this way. You should pass enableParallelism argument as True in the configuration. Its default value is False.

if __name__ == '__main__':
   config = {'algorithm': 'C4.5', 'enableParallelism': True}
   model = chef.fit(df, config)

Notice that you have to locate training step in an if block and it should check you are in main.

E-Learning

This playlist guides you how to use Chefboost step by step for different algorithms.

You can also find the tutorials about these core algorithms here.

Besides, you can enroll this online course - Decision Trees for Machine Learning From Scratch and follow the curriculum if you wonder the theory of decision trees and how this framework is developed.

Support

There are many ways to support a project - starring⭐️ the GitHub repos is just one.

Citation

Please cite chefboost in your publications if it helps your research. Here is an example BibTeX entry:

@misc{serengil2019chefboost,
  abstract = {Lightweight Decision Trees Framework supporting Gradient Boosting (GBDT, GBRT, GBM), Random Forest and Adaboost w/categorical features support for Python},
  author={Serengil, Sefik Ilkin},
  title={chefboost},
  url={https://github.com/serengil/chefboost}
  year={2019}
}

Licence

Chefboost is licensed under the MIT License - see LICENSE for more details.

Chefboost logo is created by Tang Ge. Licensed under Creative Commons: By Attribution 3.0 License.