In machine learning, the task of classification is that of dividing a set of observations (objects) into groups called classes, based on an analysis of their formal description. For classification, each observation (object) is mapped to a certain group or nominal category based on a certain qualitative property. Classification is a supervised task because it requires known classes for training samples. Labeling of a training set is usually done manually, with the involvement of specialists in the given field of study. It's also notable that if classes are not initially defined, then there will be a problem with clustering. Furthermore, in the classification task, there may be more than two classes (multi-class), and each of the objects may belong to more than one class (intersecting).

In this chapter, we will discuss various approaches to solving a classification task with machine learning. We are going to look at some of the most well-known and widespread algorithms that are logistic regression, Support Vector Machine (SVM), and k-Nearest Neighbors (kNNs). Logistic regression is one of the most straightforward algorithms based on linear regression and a special loss function. SVM is based on a concept of support vectors that helps to build a decision boundary to separate data. This approach can be effectively used with high-dimensional data. kNN has a simple implementation algorithm that uses the idea of data compactness. Also, we will show how the multi-class classification problem can be solved with the algorithms mentioned previously. We will implement program examples to see how to use these algorithms to solve the classification task with different C++ libraries.

The following topics are covered in this chapter:

• An overview of classification methods
• Exploring various classification methods
• Examples of using C++ libraries for dealing with the classification task