ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Science, Statistics and Ontology

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "doug foxvog" <doug@xxxxxxxxxx>
Date: Tue, 15 Nov 2011 15:13:25 -0500 (EST)
Message-id: <61815.129.6.59.206.1321388005.squirrel@xxxxxxxxxxxxxx>
On Mon, November 14, 2011 17:35, Rich Cooper said:    (01)

> Comments below,
> -Rich
>
> Sincerely,
> Rich Cooper
> EnglishLogicKernel.com
> Rich AT EnglishLogicKernel DOT com
> 9 4 9 \ 5 2 5 - 5 7 1 2    (02)


> doug foxvog said:
> On Mon, November 14, 2011 3:37, Matthew West said:    (03)

> > When I hit a nail  with a hammer 100 times and notice that it this
> > makes it enter the piece of wood, this is not a matter merely of
> > statistics, which we can interpret as we wish.    (04)

> > There is an underlying mechanism that we can observe.    (05)

> We observe events which are the results of
> the mechanism plus the whole context of the situation.  How we
> generalize from the observations to
> a set of induced rules is not given by the
> statistics, but by our understanding of the situation and what
> types of generalizations are valid or not.    (06)


> RC
> Agreed, and clearly stated.    (07)

> df
> Induction, not deduction, allows us to determine rules from our
> observations.  Depending upon the variation in our observations,
> the induced rules could be more strongly or more weakly supported.    (08)

> If i hit a single nail 100 times with the same hammer into the same
> spot in the same piece of wood, that does not guarantee if i hit the
> nail the 101st time it won't bend and/or break from the repeated stress.    (09)

> The situation provides evidence as to what might happen in a similar
> situation, but does not define what is similar.  No statistical
> evidence provides support for concluding that the use of another nail,
> hammer, spot, piece of wood, or person wielding the hammer would cause
> that nail to enter that piece of wood if given a blow by that person
> and hammer.  Generalization may lead to correct or incorrect rules.    (010)

> Correct in what sense?    (011)

Whether it is a valid model of the world for which it was created
to be a partial model in that deductions derived from the firing
of that rule would be consistent with observations.    (012)

> I can write a program that
> will only generate logically valid rules, but
> which rules are better supported (or supported at
> all) by the history of example data is a different
> question.  So the term "correct" can refer to
> logical correctness, which is relatively easy to
> constrain through careful use of FOL algebra, or
> "correct" can refer to whether it matches the
> actually observed data.    (013)

I was discussing inferring rules for deducing what might happen
"in a similar situation".  Therefore, i was discussing rules that
not only match the actually observed data, but that can be used
to accurately (or "correctly") conclude what would happen in future
circumstances which match the antecedent conditions in the rule --
the conjunction of which defines what the rule maker considers to
be a "similar" situation.    (014)

Not only that, but the rules to be generated would predict what
would happen to the nail in a situation that matched the rule
conditions.  A rule that "If i hit the nail with a hammer once,
i will hit it with the same hammer again" might match the observed
behaviour, but that is not the type of rule this discussion was
referring to.    (015)

> RC
> The second type is what cannot be guaranteed until
> after the newly generated rule is run past the history of data.    (016)

A rule based on data needs to be run past new data to give one an idea
of how well it models the section of reality that it is intended to model.
Seeing whether it matches the data it was designed to match says nothing
about its predictive powers, merely that it was not written consistent
with the stated versions of the known facts.    (017)

The validity of the rule is not "guaranteed" if it merely agrees with
the data upon which it was based.  It is also not "guaranteed" if it
predicts correctly what happens the next 10 times the situation matches
the rule antecedent.  New supporting evidence will only provide supporting
evidence for -- not guarantee -- the rule.    (018)

-- doug foxvog
> ...
>       -- doug foxvog
>
>       > Regards
>
>       > Matthew West
>
> Thanks for the inputs,
> -Rich    (019)

>       > Information  Junction
>       > Tel: +44 1489 880185
>       > Mobile: +44 750 3385279
>       > Skype: dr.matthew.west
>       > matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx
>       > http://www.informationjunction.co.uk/
>       > http://www.matthew-west.org.uk/    (020)

>       > This email originates from Information
> Junction Ltd. Registered in England
>       > and Wales No. 6632177.
>       > Registered office: 2 Brookside, Meadow
> Way, Letchworth Garden City,
>       > Hertfordshire, SG6 3JE.
> ...    (021)

=============================================================
doug foxvog    doug@xxxxxxxxxx   http://ProgressiveAustin.org    (022)

"I speak as an American to the leaders of my own nation. The great
initiative in this war is ours. The initiative to stop it must be ours."
    - Dr. Martin Luther King Jr.
=============================================================    (023)


_________________________________________________________________
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    (024)

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