Methods for Complex Systems Science: Statistics and Dimensionality Reduction

Catalog description

Because complex systems involve a large number of interacting components, observational studies of these systems typically generate data sets of high dimension. Examples include the hundreds or thousands of distinct neurons in a brain, genes in a cell, people in a society, firms in an economy, or texts in a corpus. To make sense of such data, a diverse set of data analytic tools has been developed to summarize key properties at the population level (statistics) and characterize predictable lower-dimensional patterns (dimensionality reduction). Provides a guided tour of such tools most relevant to complex adaptive systems. With a solid foundation in inferential statistics, students encounter PCA-type linear projections, nonlinear manifold techniques, topic modeling, clustering methods, network statistics, as well as more abstract foundations for how these methods work and when they fail. Students hone their data skills by applying state-of-the-art open-source software to real-world datasets.