Publisher's Synopsis
This book develops Multidimensinal Scaling and Dimension Reduction Methods for work in Statistics and Data Science. In addition, the book also develops examples and applications relating to such methods. Multidimensional scaling (MDS) is a set of methods that allows you to visualize how near points are to each other for many kinds of distance or dissimilarity metrics and can produce a representation of your data in a small number of dimensions. MDS does not require raw data, but only a matrix of pairwise distances or dissimilarities. multivariate statistical methods often begin with some type of dimension reduction, in which data are approximated by points in a lower-dimensional space. Dimension reduction is the goal of the methods presented in this book. Dimension reduction often leads to simpler models and fewer measured variables, with consequent benefits when measurements are expensive and visualization is important. The most important content in this book is the following: - "Multidimensional Scaling" - "Nonclassical and Nonmetric Multidimensional Scaling" - "Classical Multidimensional Scaling" - "Example: Multidimensional Scaling" - "Procrustes Analysis" - "Compare Handwritten Shapes Using Procrustes Analysis" - "Feature Selection" - "Select Subset of Features with Comparative Predictive Power" - "Feature Transformation" - "Nonnegative Matrix Factorization" - "Perform Nonnegative Matrix Factorization" - "Principal Component Analysis (PCA)" - "Analyze Quality of Life in U.S. Cities Using PCA" - "Factor Analysis" - "Analyze Stock Prices Using Factor Analysis" - "Robust Feature Selection Using NCA for Regression" - "Neighborhood Component Analysis (NCA) Feature Selection" - "t-SNE" - "t-SNE Output Function" - "Visualize High-Dimensional Data Using t-SNE" - "tsne Settings" - "Feature Extraction" - "Feature Extraction Workflow" - "Extract Mixed Signals"