W3C home > Mailing lists > Public > semantic-web@w3.org > June 2015

Re: Semantic Business Intelligence

From: Sebastian Samaruga <cognescent@gmail.com>
Date: Mon, 29 Jun 2015 16:05:41 -0300
Message-ID: <CAFnmbpUXj_Yd+Pi2OP92VKJQf4m019=EHWDtT4Monxa4+Jza_Q@mail.gmail.com>
To: Neil McNaughton <neilmcn@oilit.com>
Cc: "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>
Neil, my work is ongoing and very immature. I've just dumped a couple of
concepts of parts of what could built a Semantic backed BI framework
(discovery & mining).

And yes, i meant *prime* numbers. It's easy to grab them with the Java
BigInteger API. And I'll be using them to characterize Resources and their
relationships using factoring of their IDs. The static part of the ID is
the classID (prime PK) and the dynamic part (instanceID) is augmented by
the product of the other Resources in each Statement (Link). The main
purpose for this arrangement was initially indexing Resources and retrieve
similar Resources (by their factors). But now I wrapped all similarity
logic into 'Mappings' which essentially wraps a Resource and extracts all
of its metadata.

And Resources have classes fundamentally in two ways: 1) Enumerate the
class set by extension (its Subjects) or 2) Characterize a set of Subjects
belonging to the same class because of them sharing the same predicates.
For example, Subjects sharing 'brand', 'price', 'discount' predicates are
into the type/class 'Product'.

Thanks for taking the time for reading and replying me. Best regards,
Sebastian.


On Mon, Jun 29, 2015 at 4:29 AM, Neil McNaughton <neilmcn@oilit.com> wrote:

>  Sebastian
>
>
>
> A couple of quick comments..
>
>
>
> You say “classID: This is a numeric ID assigned to all Resources at
> creation time coming from a prime number sequence unique to all Resources…”
>
> I don’t think you really mean a *prime* number!
>
>
>
> Also you define some stuff in terms of class this and that – but what is a
> class?
>
>
>
> Best regards
>
> Neil McNaughton
>
> Editor and Publisher, Oil IT Journal
>
> Oil IT Journal is published by The Data Room SARL
>
> 7 Rue des Verrieres
>
> 92310 Sevres, France
>
> Cell - +336 7271 2642
>
> Tel - +331 4623 9596
>
> UK - +44 20 7193 1489
>
> USA - +1 281 968 0752
>
> info@oilit.com/http://www.oilit.com
>
> [image: LinkedIn] <https://www.linkedin.com/pub/neil-mcnaughton/6/498/243>
>
>
>
>
>
>
>
> *From:* Sebastian Samaruga [mailto:cognescent@gmail.com]
> *Sent:* Friday, June 26, 2015 8:53 PM
> *To:* semantic-web@w3.org; pragmaticweb@lists.spline.inf.fu-berlin.de;
> topicmapmail@infoloom.com; fca-list@cs.uni-kassel.de;
> sw-announce@semanticplanet.com; data-mining@googlegroups.com
> *Subject:* Semantic Business Intelligence
>
>
>
> Hi all. There is a draft I've been working for a while and I'm needing
> help trying to laid down some ideas for the core concepts of a platform
> that should enable data/schema merging, discovery and augmentation of data
> provided by many different sources.
>
> Then there should be an API exploiting the concept of 'arithmetic'
> inference and reasoning using 'meaningful' numeric identifiers. The
> document describing the ongoing outermost concepts is in a PDF file
> attached to this document. If there are problems with the attachment the
> file is accessible at:
>
> https://drive.google.com/open?id=0BxxuOINjaiBNcW11ODJXYkFzUGs&authuser=0
>
> Some techniques on Mappings similarity hashing and merging and Template
> aggregation for presentation/layers interaction purposes left to be defined
> completely.
>
> The most important thing I'm needing right now are comments and criticisms
> that could help me refine my concepts. Thanks in advance,
>
>
>
> Best regards,
>
> Sebastian Samaruga.
>



image001.png
(image/png attachment: image001.png)

Received on Monday, 29 June 2015 19:08:19 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:43:00 UTC