A data scientist is working on a regression model and wants to prevent overfitting while maintaining the model's predictive accuracy. Which of the following regularization techniques should they choose to apply?

Question

Seekh · Accepted Answer

The data scientist should use L2 regularization, also called ridge regression. L2 adds a penalty proportional to the square of the model coefficients, which shrinks all weights toward zero but keeps them non‑zero, preserving the overall pattern learned. This reduces variance by discouraging overly large weights that would fit noise, while still allowing the model to capture the underlying signal. For example, if a linear model learns a coefficient of 10 on a noisy feature, ridge would reduce it to about 8, keeping the feature useful but less extreme. Thus, L2 regularization strikes a good balance between preventing overfitting and maintaining predictive power.

A data scientist is working on a regression model and wants to prevent overfitting while maintaining the model's predictive accuracy. Which of the following regularization techniques should they choose to apply?

Learning Path

Choose the Best Answer

Understanding the Answer

Key Concepts

Deep Dive: Regularizers in Predictive Models

Definition

Topic Definition

Ready to Master More Topics?