Question

data set

from sklearn.datasets import load_diabetes
diabetes = load_diabetes()
from sklearn.model_selection import train_test_split
diabetes_X_train, diabetes_X_test, diabetes_y_train, diabetes_y_test = train_test_split(diabetes['data'], diabetes['target'], random_state=0)

What is the training and test R2 for the Lasso model using the default parameters? How many features does this model use? What are the names of those features?

Answer 1

Here is the code to fit model:

from sklearn.linear_model import Lasso

lr = Lasso()

lr.fit(diabetes_X_train, diabetes_y_train)

Code to compute R-squared on train and test:

from sklearn.metrics import r2_score

tr_p = lr.predict(diabetes_X_train)

print(r2_score(diabetes_y_train, tr_p))

Score on Train: 0.4141

te_p = lr.predict(diabetes_X_test)

print(r2_score(diabetes_y_test, te_p))

Score on Test: 0.2781

We build model on all the features but since lasso penalizes the coefficients and reduces to 0, there are only 2 featires left:

1.) bmi

2.) s5