This chapter aims to explain some dimensionality reduction techniques in machine learning. This concept refers to the process of reducing the number of random variables considered and can be subdivided in feature selection and feature extraction. The key is to reduce the number of dimensions, while preserving most parts of the information.
In this chapter, we will cover feature extraction techniques, while feature selection will be approached in the next chapter.
These kinds of techniques aim to solve the problem of dimensionality, which refers to the inconveniences associated with multivariate data analysis when the dimensionality, that is the number of variables, is too large.
The problem of dimensionality implies ...