Which term describes a model with high bias but low variance?

Which term describes a model with high bias but low variance?

• A. Underfitted Models
• B. Overfitted models
• C. Ridge Regression (L2 penalty)
• D. Elastic Net (L1 + L2 penalties)

Answer: (A)

High bias with low variance happens when the model is too simple to capture the underlying patterns in the data. It consistently makes systematic errors across different training samples because it can’t fit the data well enough. That situation is called underfitting. It contrasts with overfitting, where the model fits training data very well (low bias) but varies a lot with different data sets (high variance). Regularization methods like ridge regression and elasticNet are used to reduce variance and improve generalization, often at the cost of a bit more bias, but they don’t produce the high bias with low variance pattern. So the term that best describes a model with high bias but low variance is underfitting.

High bias with low variance happens when the model is too simple to capture the underlying patterns in the data. It consistently makes systematic errors across different training samples because it can’t fit the data well enough. That situation is called underfitting. It contrasts with overfitting, where the model fits training data very well (low bias) but varies a lot with different data sets (high variance). Regularization methods like ridge regression and elasticNet are used to reduce variance and improve generalization, often at the cost of a bit more bias, but they don’t produce the high bias with low variance pattern. So the term that best describes a model with high bias but low variance is underfitting.