A startup is developing a predictive model to forecast sales based on various marketing strategies. They decide to use a linear regression model as their parametrized predictor. Which of the following statements best describes a crucial aspect of their model design related to the parametrized nature of the predictor?

A linear regression model learns a set of coefficients that describe how each marketing variable shifts the expected sales; these coefficients are the model’s parameters. Because the model is parametrized, the quality of the forecast depends on how well those parameters are estimated from the training data. To avoid over‑fitting, the startup must choose a sensible set of predictors, scale them appropriately, and validate the model on unseen data. For example, if the model assigns a coefficient of 0. 5 to email‑campaign spend, it predicts that each additional dollar spent raises sales by fifty cents, assuming the relationship holds.