In linear regression, all independent variables must be numeric.

• A. Numeric
• B. Categorical
• C. Binary
• D. Text

Answer: (A)

Linear regression models the target as a weighted sum of the predictors: y = beta0 + beta1x1 + ... + betapxp. For this to work, each predictor must be numeric so you can multiply it by its coefficient and add it to the rest of the terms. Categorical predictors are not numbers in their raw form, so they can’t enter the model directly; they must be converted into numeric indicators through encoding (for example, one-hot). Binary variables are naturally numeric (0 or 1) and fit directly as predictors. Text data isn’t numeric either and needs to be transformed into numeric features via vectorization before inclusion. Because the question asks for the general requirement, numeric predictors is the correct framing, making the option Numeric the best choice.