Classification is a supervised machine learning task where the goal is to assign a discrete class (or category) to each data point. For example: handwritten digit recognition is a classification problem because each digit belongs to one of 10 distinct classes.