Which of the following best describes the role of coefficients in linear regression?

• A. They determine predicted class.
• B. They represent the relationship strength between inputs and the target.
• C. They are always non-negative.
• D. They are used to compute probabilities.

Answer: (B)

In linear regression, the coefficients are the weights that scale each input feature to its contribution to the predicted value. They show how strongly and in which direction the target changes as a feature varies, holding other features constant. A positive coefficient means the prediction goes up with that feature; a negative one means it goes down. The larger the magnitude, the stronger the influence of that feature on the prediction. This is why coefficients represent the relationship strength between inputs and the target.

Predicted class is tied to classification, not regression, so coefficients in linear regression aren’t about determining a class. Coefficients can be negative or positive, so they aren’t required to be non-negative. Probabilities aren’t directly computed in standard linear regression; probability estimates arise in models like logistic regression or other probabilistic frameworks.