Modern Dimension Reduction

Data are increasingly complex, and dimension reduction techniques offer a way to simplify them. This element provides readers with a suite of modern unsupervised dimension reduction techniques and hundreds of lines of R code to represent high-dimensional data in a lower-dimensional subspace.

Format: Paperback / softback
Length: 75 pages
Publication date: 05 August 2021
Publisher: Cambridge University Press

Data have become an integral part of our society, and their complexity is growing rapidly in both size and dimensionality. Dimension reduction techniques provide researchers and scholars with the ability to simplify and manage these complex data spaces. This Element offers readers a comprehensive guide to modern unsupervised dimension reduction methods, along with hundreds of lines of R code, to effectively represent high-dimensional data in a simplified, lower-dimensional subspace.

We begin by exploring the foundational technique of principal components analysis (PCA), which is widely used in dimension reduction. We then move on to introduce and demonstrate the application of several advanced techniques, including locally linear embedding (LLE), t-distributed stochastic neighbor embedding (t-SNE), uniform manifold approximation and projection (UMAP), self-organizing maps (SOM), and deep autoencoders (DAE).

By leveraging these unsupervised algorithms, readers will gain a well-stocked toolbox of techniques to tackle the complexities of high-dimensional data, which are prevalent in modern society. All code used in this Element is publicly accessible on GitHub, allowing readers to further explore and apply these methods to their own data sets.

In conclusion, dimension reduction plays a crucial role in analyzing and understanding complex data in our society. This Element provides readers with a comprehensive guide to modern unsupervised dimension reduction techniques, along with practical examples and code, to facilitate their research and analysis endeavors.

Weight: 160g
Dimension: 151 x 229 x 11 (mm)
ISBN-13: 9781108986892