Dimensionality reduction

What is dimensionality reduction?

Dimensionality analysis is a task of unsupervised machine learning concentrated on transforming useful data from a high-dimensional space into a low-dimensional space keeping the most important properties of the original data. Dimensionality reduction is often used to simplify the further analysis process and data visualization. Dimensionality reduction is widespread in fields dealing with large amounts of data, particularly signal processing, natural language processing, and computer vision.

Why is it important?

Dimensionality reduction is applicable when you have too many features to operate with. It allows you to decrease the number of variables while keeping the most critical properties. In many cases, dimensionality reduction is used for visualization purposes because we cannot graphically represent more than three variables at once.