gradientcobra v1.0.8
package
Installing and importing packages
gradientcobra can be installed from pypi using pip
:
pip install gradientcobra
The package contains the following methods:
Gradient COBRA
Gradient COBAR is a kernel-based consensual aggregation method for regressoin (S. Has (2023)). After installing gradientcobra
package (version \(\geq\) v1.0.6
), GradientCOBRA
method can be imported by:
from gradientcobra.gradientcobra import GradientCOBRA
For more information about this method, read GradientCOBRA documentation.
MixCOBRARegressor
MixCOBRA is an aggregation method for both regression and classification that takes into account both input and output of the basic estimators (A. Fischer and M. Mougeot (2019)). From version 1.0.6
, this method is also available for regression and can be implemented in python by:
from gradientcobra.mixcobra import MixCOBRARegressor
A challenge of this method is that it has to be optimized over a \(2\)D-grid of parameters \((\alpha, \beta)\) for input and output part. In gradientcobra
library, the implementation of this method is optimized so that it works very efficiently. Moreover, you can also use gradient descent algorithm to estimate the parameters \((\alpha, \beta)\), or only one bandwidth parameter \(h\) can be used for both input and output part. For more information on how to use this method, read MixCOBRARegressor documentation.
References
- S. Has (2023). A Gradient COBRA: A kernel-based consensual aggregation for regression. Journal of Data Science, Statistics, and Visualisation, 3(2).
- A. Fischer, M. Mougeot (2019). Aggregation using input-output trade-off. Journal of Statistical Planning and Inference, 200.
- G. Biau, A. Fischer, B. Guedj and J. D. Malley (2016), COBRA: A combined regression strategy, Journal of Multivariate Analysis.
- M. Mojirsheibani (1999), Combining Classifiers via Discretization, Journal of the American Statistical Association.