Straight or Curved? — medium Machine Learning in Code problem | Incognition

Straight or Curved?

10 pts · 10 coins

Machine Learning in CodeMedium

Straight or Curved?

A clinic can only afford some of the ten diabetes measurements. For a given set of columns, they want to know whether a curved model earns its extra columns.

Write straight_or_curved(columns).

Load the data with load_diabetes(return_X_y=True, as_frame=True) and split it with train_test_split(X, y, test_size=0.25, random_state=0).
Keep only the columns named in columns.
Straight: make_pipeline(StandardScaler(), LinearRegression()).
Curved: make_pipeline(PolynomialFeatures(2), StandardScaler(), LinearRegression()).
Fit both on the training rows and score both on the test rows (R²).

Return a dict:

"straight": the straight model's test R², rounded to 3 places.
"curved": the curved model's test R², rounded to 3 places.
"ship": "curved" if the curved model's unrounded test R² is higher, otherwise "straight".

Examples

columns = ['s3']

returns {'straight': 0.064, 'curved': 0.076, 'ship': 'curved'}

columns = ['bmi', 's5']

returns {'straight': 0.276, 'curved': 0.261, 'ship': 'straight'}

Hints

solution.py

1from sklearn.datasets import load_diabetes
2from sklearn.linear_model import LinearRegression
3from sklearn.model_selection import train_test_split
4from sklearn.pipeline import make_pipeline
5from sklearn.preprocessing import PolynomialFeatures, StandardScaler
6
7
8def straight_or_curved(columns):
9    pass
10