Re: [xml-dev] Schemaless XML?

On Tue, Oct 11, 2016 at 8:53 AM, Costello, Roger L. <costello@mitre.org> wrote:

Hi Folks,

An alternate approach is for the application to go “schemaless.” The application performs machine learning on the data it receives. I’m not sure what “machine learning on the data” means. I suspect that it means that an internal schema (in some form or another) is dynamically generated. Do you agree?

No.

Machine learnings utility lies in solving problems that we either don't know how to solve or are intractable.

To illustrate. A tractable (not too big) problem in linear regression has a known closed formula, so while we could use machine learning to learn it (or rather an approximation of it) it would be rather pointless to do so.

Same here, "schemaless" or not you should know what your data means and is going to be used for.