A theorem establishing that performance on test data cannot be deduced from performance on training data. It follows that the justification for any particular learning algorithm must be based on an assumption that nature is uniform in some way. Since different machine learning algorithms make such different assumptions, no-free-lunch theorems have been used to argue that it not possible to deduce that any algorithm is superior to any other from first principles. Thus “good” algorithms are those whose inductive bias matches the way the world happens to be.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Further Reading
Wolpert DH, Macready WG (1997) No free lunch theorems for optimization. IEEE Trans Evol Comput 1(1):67–82
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer Science+Business Media New York
About this entry
Cite this entry
(2017). No-Free-Lunch Theorem. In: Sammut, C., Webb, G.I. (eds) Encyclopedia of Machine Learning and Data Mining. Springer, Boston, MA. https://doi.org/10.1007/978-1-4899-7687-1_592
Download citation
DOI: https://doi.org/10.1007/978-1-4899-7687-1_592
Published:
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4899-7685-7
Online ISBN: 978-1-4899-7687-1
eBook Packages: Computer ScienceReference Module Computer Science and Engineering