Dimensionality Reduction

Dimensionality reduction refers to several mathematical methods for reducing the number of variables under consideration. The motivation for DR is that some data analysis techniques, such as regression, classification, and even simple visualization, work better with fewer dimensions.

One can reduce dimensions by selecting the most important ones or projecting data from a higher-dimensional space into a space with fewer dimensions. Principal Component Analysis (PCA) is an example of the latter that relies on making linear combinations of variables.

Last Updated: April 07, 2019