What type of machine learning problem does a support vector machine typically solve?

A support vector machine (SVM) is primarily designed to solve classification problems. It works by finding the optimal hyperplane that best separates different classes in the data. The idea is to maximize the margin between the closest points of different classes, known as support vectors, and this maximization leads to improved classification performance on new, unseen data.

In a classification task, the SVM takes labeled training data and learns how to categorize the input features into predefined classes. When presented with new data, the SVM can effectively determine which class the new instances belong to, based on the learned hyperplane.

While support vector machines can also be adapted for regression tasks (leading to a variation known as support vector regression), their fundamental design and most common application lie within classification tasks. Choosing this option reflects an understanding of SVM's primary use case in machine learning, distinguishing it from other types of problems like unsupervised clustering or data reduction.