W3C home > Mailing lists > Public > semantic-web@w3.org > June 2013

[semantic-web] <none>

From: Sebastian Samaruga <cognescent@gmail.com>
Date: Fri, 7 Jun 2013 23:22:08 -0300
Message-ID: <CAFnmbpWzLW23EAUUD+YFaicpj2MEajyT__tsBA+fX=UUMh0+eA@mail.gmail.com>
To: "semantic-web@w3.org" <semantic-web@w3.org>, "pragmaticweb@lists.spline.inf.fu-berlin.de" <pragmaticweb@lists.spline.inf.fu-berlin.de>, "topicmapmail@infoloom.com" <topicmapmail@infoloom.com>

I'm trying not to reinvent the wheel. So I'm posting this because I need to
homogenize data from any data source for aggregation and consolidation for
a proposed BI framework. I was wondering if there are ways of doing this
without having any original source data schema related information.

So I tried to figure out how could I rebuild or learn schema information
from source data. As an example, as first I thought I could infer some
subject type because of the predicates of the statements in which it
occurs. So, for example, if someone appears as subject of a series of
(same) predicates, those subjects belong to the same type or class.

I know there are ways of doing this 'semantically'. But I needed to do this
without RDFS or OWL, at least in principle.

I'm attaching the very first draft of what I think I should need. Also
there is the link for it in Google Drive. Apologies if it is confuse. I
only liked to share the diagrams and a few ideas.



Received on Saturday, 8 June 2013 02:23:36 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:45:33 UTC