Jump to content

Talk:Learning theory (statistics)

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

I'm not sure what is meant by "mind changes" ? Is this talking about VC-dimension (or fat-shattering) bounds? -- hike395 15:41 17 Jul 2003 (UTC)

Moved from article, User:Michael Hardy suggests ---

Should this page be merged with statistical learning theory?

I would suggest not. Statistical learning theory is a specific, poorly-named, form of computational learning theory, established by Vladimir Vapnik. It turns out that the practitioners of the different brands of learning theory (e.g., Bayesian, PAC, etc.) have strong philosophical and practical disagreements with the Vapnik-style theorists, and would not agree that they are the same thing at all.

One possibility is renaming this article "Computational learning theory", which is the name of the main conference of the field. That may be clearer. What do people think? Another possibility is putting "Vapnik" into the title of "Statistical learning theory", but I am not sure how to do that without being idiosyncratic and making that other article hard to find. --- hike395 02:00, 8 Oct 2004 (UTC)