Cyclic Boosting Machines - an explainable supervised machine learning algorithm
Перейти к файлу
microsoft-github-policy-service[bot] 13d2c25320
Microsoft mandatory file (#13)
Co-authored-by: microsoft-github-policy-service[bot] <77245923+microsoft-github-policy-service[bot]@users.noreply.github.com>
2022-07-25 21:26:07 +02:00
.github/workflows added categorical, date/time and continuous feature support added feature importance plotting 2021-11-24 11:28:26 +01:00
.vscode Fancy badges (#4) 2021-09-29 00:02:44 +02:00
cbm scikit-learn compatibility (#12) 2022-05-04 17:29:10 +02:00
data initial clean code 2021-05-10 09:53:54 +02:00
images Improved plotting & docs (#10) 2021-11-29 10:08:33 +01:00
kaggle scikit-learn compatibility (#12) 2022-05-04 17:29:10 +02:00
src scikit-learn compatibility (#12) 2022-05-04 17:29:10 +02:00
tests scikit-learn compatibility (#12) 2022-05-04 17:29:10 +02:00
.gitattributes Improved plotting & docs (#10) 2021-11-29 10:08:33 +01:00
.gitignore Initial commit 2021-05-10 06:44:48 +00:00
CODE_OF_CONDUCT.md Initial CODE_OF_CONDUCT.md commit 2021-05-09 23:44:58 -07:00
LICENSE Initial LICENSE commit 2021-05-09 23:44:59 -07:00
MANIFEST.in added categorical, date/time and continuous feature support added feature importance plotting 2021-11-24 11:28:26 +01:00
README.md Improved plotting & docs (#10) 2021-11-29 10:08:33 +01:00
SECURITY.md Microsoft mandatory file (#13) 2022-07-25 21:26:07 +02:00
pyproject.toml Badge improvements (#3) 2021-09-20 18:32:06 +02:00
setup.py Openmp (#11) 2021-12-10 20:59:22 +01:00

README.md

Cyclic Boosting Machines

Build Python codecov PyPI version License: MIT Academic Paper

This is an efficient and Scikit-learn compatible implementation of the machine learning algorithm Cyclic Boosting -- an explainable supervised machine learning algorithm, specifically for predicting count-data, such as sales and demand.

Features

  • Optimized for categorical features
  • Continuous features are discretized using pandas.qcut.
  • Date auto-expansion (weekday + month).
  • Feature importance plots: categorical, continuous and interactions.
  • Metrics to stop training: RMSE, L1, SMAPE.

Usage

The CBM model predicts by multiplying the global mean with each weight estimate for each bin and feature. Thus the weights can be interpreted as % increase or decrease from the global mean. e.g. a weight of 1.2 for the bin Monday of the feature Day-of-Week can be interpreted as a 20% increase of the target.

with

pip install cyclicbm
import cbm
from sklearn.metrics import mean_squared_error

# load data using https://www.kaggle.com/c/demand-forecasting-kernels-only
train = pd.read_csv('data/train.csv', parse_dates=['date'])
test  = pd.read_csv('data/test.csv',  parse_dates=['date']) 

# feature engineering
min_date = train['date'].min()

def featurize(df):
    out = pd.DataFrame({
        # TODO: for prediction such features need separate modelling
        'seasonal' : (df['date'] - min_date).dt.days // 60,
        'store'    : df['store'], 
        'item'     : df['item'], 
        'date'     : df['date'],
        # <name-1> _X_ <name-2> to mark interaction features
        'item_X_month': df['item'].astype(str) + '_' + df['date'].dt.month.astype(str)
    })
    
    return out

x_train_df = featurize(train)
x_test_df  = featurize(test)
y_train = train['sales']

# model training
model = cbm.CBM()
model.fit(x_train_df, y_train)

# test on train error
y_pred_train = model.predict(x_train_df).flatten()
print('RMSE', mean_squared_error(y_pred_train, y_train, squared=False))

# plotting
model.plot_importance(figsize=(20, 20), continuous_features=['seasonal'])

Feature Importance Plot

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Trademarks

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.