[Top] [All Lists]

Re: [ontolog-forum] Deriving categories using clustering techniques...

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Rich Cooper" <rich@xxxxxxxxxxxxxxxxxxxxxx>
Date: Thu, 16 Jul 2009 09:15:03 -0700
Message-id: <20090716161521.55952138D1A@xxxxxxxxxxxxxxxxx>


Rich Cooper


Rich AT EnglishLogicKernel DOT com


John Bottoms wrote:

JB>  My understanding of categories consists of the following:

(domain == professional discipline or context across time

            or across a number of contributors)


RC>  If this is a continuation of the Semantic Systems thread (?), then some of the emails we have been seeing about "Measures" might be relevant to distinguishing individuals from groups.  If we need a Property (Measure), we need to collect specimen individuals using that Measure before we can categorize the individuals or the group in any way. 


But it still seems necessary to have a "window function" ("membership predicate?") that will distinguish A from Not A specimens based on the Value of the measured Property, thus creating the first Category - A. 


Every Category requires a Predicate, so that must be next. 


After that, I concede that Conjunction be canonized into the ontogeny. 






1. Quantification of an environment can be accomplished

    using categories.


2. The domain is determined by a boundary and the ideal

    quantification process divides the domain efficiently.


3. The metric for the effectiveness quantification is the

    coverage of the domain. However, this may also

    include metrics that include more functionally useful

    components such as means.


My question is related to the effective coverage and is, in

part, derived from the criticism of the DIKW (or any other

pyramid) of the abstractions of knowledge.


Q: "Should it be possible to replace the existing approach to

     categories with an approach that extracts measured concepts

     from meanings and then performs cluster analysis on those



RC>  Yes.  For one example, see:



JB> Rational:

a.) One of the weakness of the category concept is that they

     are designed for coverage fo a domain for which the boundary

     may be poorly ascribed or poorly understood. e.g. we have

     seen redefinition of "I.Q" to include spatial and musical



b.) Next, domain boundaries change across time and new categories

     may need to be added; for those changes the new entities must

     be vetted. This could be done using cluster analysis.


c.) Finally, the redefinition of the process for determining

     categories could be automated and because the process relies

     on a set of meanings rather than a set of categories, it is

     easier to perform vetting on the meanings.


RC> John that last point isn't clear to me at all, but all the preceding sounds like a great start on reviewing identification and categorization processes. 


-John Bottoms


  Concord, MA

  T: 978-505-9878


Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)

<Prev in Thread] Current Thread [Next in Thread>