Beat the Lazy Guess — medium Machine Learning in Code problem | Incognition

Beat the Lazy Guess

10 pts · 10 coins

Machine Learning in CodeMedium

Beat the Lazy Guess

Before trusting any model, compare it with the laziest one possible: always predict the average of the training answers, whatever the row.

Write lazy_baseline(train_answers, test_answers).

Work out the mean of train_answers.
Predict that same number for every value in test_answers.
Score those predictions against test_answers.

Return a dict with:

"rmse": the root mean squared error, rounded to 2 decimal places.
"r2": the R² score, rounded to 3 decimal places.

Use plain Python floats. test_answers always holds at least two different values.

Examples

train_answers = [10, 20, 30]

test_answers = [15, 25]

returns {'rmse': 5.0, 'r2': 0.0}

train_answers = [12, 14, 16, 18]

test_answers = [10.0, 12.0, 14.0, 16.0]

returns {'rmse': 3.0, 'r2': -0.8}

Hints

solution.py

1import numpy as np
2from sklearn.metrics import mean_squared_error, r2_score
3
4
5def lazy_baseline(train_answers, test_answers):
6    pass
7