Fewer Measurements, Lower Score?

Some blood tests are expensive. A clinic wants to know how well a model does if it only uses a few of the ten measurements.

Write test_score_with(columns).

• Load the data with load_diabetes(return_X_y=True, as_frame=True).
• Split it with train_test_split(X, y, test_size=0.25, random_state=0).
• Keep only the columns named in columns, in both the training and test rows.
• Fit a LinearRegression model on the training rows.
• Return its R² score on the test rows, rounded to 3 decimal places.