Dimensionality reduction is the process of transforming a high-dimensional data set, or one with many features, into a more compact representation with fewer features. Reducing the dimension of a data set can be advantageous for computation time as well as preprocessing for supervised learning algorithms.