Metadata-Version: 2.1
Name: gradientcobra
Version: 0.0.3
Summary: Python implementation for Gradient COBRA: A kernel-based consensual aggregation for regression by S. Has (2023).
Home-page: https://github.com/hassothea/gradientcobra/
Author: ['Sothea Has']
Author-email: sothea.has@lpsm.paris
Keywords: Consensual aggregation,Kernel,Regression,Statistical Aggregation
Classifier: Development Status :: 4 - Beta
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python
Classifier: Operating System :: Microsoft :: Windows
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering
License-File: LICENSE
Requires-Dist: numpy>=1.20.1
Requires-Dist: pandas>=2.1.0
Requires-Dist: scipy>=1.10.1
Requires-Dist: scikit-learn>=1.2
Requires-Dist: matplotlib
Requires-Dist: seaborn

# gradientcobra

This is the `python` package implementation of `Gradient COBRA` method by [S. Has (2023)](https://jdssv.org/index.php/jdssv/article/view/70). 

## Summary

Gradient COBRA is a kernel-based consensual aggregation for regression problem that extends the naive kernel-based of Biau et al. (2016) to a more general regular kernel-based configuration. It is theoretically shown that Gradient COBRA inherits the consistency of the consistent basic estimator in the list, and the same rate of convergence is archived for exponentially tail-delaying kernel functions. On top of that, gradient descent algorithm is proposed to efficiently estimate the bandwidth parameter of the aggregation method. It is shown to be up to hundred times faster than the classical method and `python` package [pycobra](https://arxiv.org/abs/1707.00558).
