MLS-C01 Question #304: Real Exam Question with Answer & Explanation

Question

An exercise analytics company wants to predict running speeds for its customers by using a dataset that contains multiple health-related features for each customer. Some of the features originate from sensors that provide extremely noisy values. The company is training a regression model by using the built-in Amazon SageMaker linear learner algorithm to predict the running speeds. While the company is training the model, a data scientist observes that the training loss decreases to almost zero, but validation loss increases. Which technique should the data scientist use to optimally fit the model?

Options

• AAdd L1 regularization to the linear learner regression model.
• BPerform a principal component analysis (PCA) on the dataset. Use the linear learner regression
• CPerform feature engineering by including quadratic and cubic terms. Train the linear learner
• DAdd L2 regularization to the linear learner regression model.

Explanation

{"question_number": 4, "question_summary": "Fix overfitting (training loss ≈ 0, validation loss rising) in a linear learner regression model trained on noisy sensor data.", "correct_answer": "A", "explanation": "The pattern of near-zero training loss with rising validation loss is the textbook signature of overfitting. L1 regularization (Lasso) is the best fit here because it penalizes the absolute value of coefficients, which drives the weights of irrelevant or noisy features all the way to zero. This directly addresses the described problem: noisy sensor features contribute spurious signal during training that doesn't generalize. L2 regularization (D, Ridge) shrinks all coefficients proportionally but rarely zeroes them out, so noisy features remain active and the overfitting may persist. Option B (PCA) can reduce noise but adds preprocessing complexity. Option C (adding polynomial terms) would increase model complexity and worsen overfitting.", "generated_by": "claude-sonnet", "llm_judge_score": 3}

No community discussion yet for this question.