Getting Started’s Python module bbai automatically tunes hyperparameters for logistic regression, ridge regression, and other GLMs to optimize Approximate Leave-one-out Cross-validation.

To install bbai for Mac or Linux, run

pip install bbai

Fitting Ridge Regression

We’ll use bbai to find the hyperparameters for ridge regression that optimize Leave-one-out Cross-validation.

First, load an example data set.

from sklearn.datasets import load_boston
from sklearn.preprocessing import StandardScaler
X, y = load_boston(return_X_y=True)
X = StandardScaler().fit_transform(X)

Next, fit the model. Because bbai uses second-order information to find the best hyperparameters, there’s no need to specify a search space.

from bbai.glm import RidgeRegression
model = RidgeRegression(), y)

We can now print out the hyperparameter found.

print('alpha = ', model.alpha_)


alpha =  4.680170622758263

Fitting Logistic Regression

We can also find hyperparameters for logistic regression that optimize Approximate Leave-one-out Cross-validation.

Load an example classification data set.

from sklearn.datasets import load_breast_cancer
from sklearn.preprocessing import StandardScaler
X, y = load_breast_cancer(return_X_y=True)
X = StandardScaler().fit_transform(X)

Fit and find the best hyperparameters.

from bbai.glm import LogisticRegression
model = LogisticRegression(), y)

Print out the hyperparameter we found.

print('C = ', model.C_)


C =  0.6655139682151202