Linear regression is suitable for predicting continuous values, not categories. Which term describes this limitation?

Linear regression is suitable for predicting continuous values, not categories. Which term describes this limitation?

• A. Dummy Variables
• B. Intercept
• C. Inappropriateness for Classification
• D. Non-Linear Terms

Answer: (C)

Linear regression is designed to predict a continuous outcome, not discrete categories. The term that best describes this limitation is the inappropriateness for classification. Since the model outputs real-valued predictions, using it directly for class labels would require arbitrary thresholds to convert those predictions into categories, which can lead to poor and unstable decisions. Classification-focused methods, like logistic regression or other classifiers, are built to handle discrete outcomes by modeling the probability of class membership and making decisions based on that probability. The other options don’t capture this fundamental mismatch: dummy variables are just a way to encode categorical inputs, the intercept is the base value of the prediction, and non-linear terms expand the model to capture curvature but don’t address the issue of predicting categories.

Linear regression is designed to predict a continuous outcome, not discrete categories. The term that best describes this limitation is the inappropriateness for classification. Since the model outputs real-valued predictions, using it directly for class labels would require arbitrary thresholds to convert those predictions into categories, which can lead to poor and unstable decisions. Classification-focused methods, like logistic regression or other classifiers, are built to handle discrete outcomes by modeling the probability of class membership and making decisions based on that probability.

The other options don’t capture this fundamental mismatch: dummy variables are just a way to encode categorical inputs, the intercept is the base value of the prediction, and non-linear terms expand the model to capture curvature but don’t address the issue of predicting categories.