How do classification tasks differ from regression tasks?

Classification tasks are fundamentally designed to predict categorical outcomes, meaning they classify data points into discrete groups or categories. For example, a classification model may predict whether an email is 'spam' or 'not spam' based on certain features of the email. This is in contrast to regression tasks, which focus on predicting continuous values.

In a regression task, the output can be any real number, allowing for a wide range of possible predictions, such as predicting house prices based on various features like size and location. The core difference lies in the nature of the outputs: classification deals with distinct labels, while regression concerns itself with numeric values across a continuous spectrum.

Additionally, the other choices highlight misunderstandings of the definitions of classification and regression. They incorrectly associate classification with continuous outcomes, and regression with categorical outcomes and unstructured data, which do not accurately represent the characteristics of these tasks.