In machine learning, what type of classification problem allows data examples to be classified into one of three or more classes?

Multi-class classification is the correct classification problem type where data examples can be classified into one of three or more classes. This approach is fundamental in machine learning, especially when the target variable can take on multiple discrete values.

In multi-class classification, each instance in the dataset is assigned to one and only one class from a set of more than two classes. For example, classifying types of fruits into categories such as apples, oranges, and bananas represents a typical multi-class scenario.

This differs from binary classification, where there are only two possible classes for each example, such as spam or not spam. Multi-label classification involves assigning multiple classes to a single instance, which means one data point can belong to various categories at the same time. Regression classification, while sounding similar, refers mostly to predicting continuous values rather than categorizing instances into classes.

Understanding these distinctions is crucial for effectively framing machine learning problems, as the choice between multi-class, binary, and other classification types impacts the algorithms used, the evaluation metrics, and ultimately the model's performance.