Back to search

Cornell University

Science In High Dimensions: Multiparameter Models And Big Data

Abstract

dc:description.abstract

Complex multiparameter models such as in climate science, economics, systems biology, materials science, neural networks and machine learning have a large-dimensional space of undetermined parameters as well as a large-dimensional space of predicted data. These high-dimensional spaces of inputs and outputs pose many challenges. Recent work with a diversity of nonlinear predictive models, microscopic models in physics, and analysis of large datasets, has led to important insights. In particular, it was shown that nonlinear fits to data in a variety of multiparameter models largely rely on only a few stiff directions in parameter space. Chapter 2 explores a qualitative basis for this compression of parameter space using a model nonlinear system with two time scales. A systematic separation of scales is shown to correspond to an increasing insensitivity of parameter space directions that only affect the fast dynamics. Chapter 3 shows with the help of microscopic physics models that emergent theories in physics also rely on a sloppy compression of the parameter space where macroscopically relevant variables form the stiff directions. Lastly, in chapter 4, we will learn that the data space of historical daily stock returns of US public companies has an emergent simplex structure that makes it amenable to a low-dimensional representation. This leads to insights into the performance of various business sectors, the decomposition of firms into emergent sectors, and the evolution of firm characteristics in time.

Degree

thesis:*
Name thesis:degree_name
Ph. D., Theoretical and Applied Mechanics
Level thesis:degree_level
Doctor of Philosophy
Discipline thesis:degree_discipline
Theoretical and Applied Mechanics
Grantor
Cornell University
Year dc:date.issued
2014

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Chachra, Ricky
Chair dc:contributor.chair
  • Sethna, James Patarasp
Committee members dc:contributor.committeemember
  • Strogatz, Steven H
  • Guckenheimer, John Mark

Subjects

dc:subject × 3

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Dc Identifier Bibid
8442347

Chain of custody

source
Harvested from
Cornell University
Base URL
ecommons.cornell.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Chachra, Ricky. Science In High Dimensions: Multiparameter Models And Big Data. Doctor of Philosophy thesis, Cornell University, 2014. https://hdl.handle.net/1813/36155