Transforming a dataset into one with fewer columns is more complicated than it might seem, explains Dr. James McCaffrey of Microsoft Research in this full-code, step-by-step machine learning tutorial.

We examined the ability of eigenvalue tests to distinguish field-collected from random, assemblage structure data sets. Eight published time series of species abundances were used in the analysis, ...