Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: Inference search engines? -- links to inference searches included.

__/ [Noticedtrends] on Thursday 13 October 2005 21:50 \__

> Can inference search-engines narrow-down the number of often irrelevant
> results, by using specific keywords; for the purpose of discerning
> emerging social & business trends?

Yes, but I guess it was rhetorical or maybe the question was intended to ask
if engines 'in the wild' are capable or are doing this.

> For example, if authors of content subconsciously mention the keywords
> "lately and noticed" within the same brief sentence, the reader may
> infer "an unintended message" through the process of "inferential
> scanning;" a method of noting "semantic anomalies" which may
> signal emerging trends.

I have been thinking about such concepts when registering http://iuron.com a
few days ago. There is not much use of any semantics, yet.

> Note: The wildcard character asterisk "*" is important for two
> examples:
> -- In the current month and year, the wildcard character "*"
> represents a specific date. In quotations: "October * 2005" infers
> content that mentions the current month and year. Yet, search results
> also include older content.

The problem with much of the content on the Web is that even dates are
hard-coded in unnatural ways. As oppose to XML, standardised metadata or
the like, dates are embedded as standard content, unlike LaTeX, for

> -- Search terms in quotations with the wildcard character "*"
> indicates two words NEAR each other: "lately * noticed." This type
> of search is excellent for trend spotting.

What would be the intended practicality? Use tags in Technorati, for
example, to spot trends. Tagclouds come to mind too. They feeds on plenty
of fresh RSS'd (XML, hence time labelled) content.

> -- Search results are further narrowed-down by an astounding 80-90% by
> excluding such words he, she, me. After all, content which may signal
> emerging trends usually doesn't contain 'he, she, me.' Certain
> pronouns usually indicate personal accounts of individuals; rather than
> the "collective conscious" of people & businesses developing into
> emerging trends.
> -- Examples of links offer plenty of ideas for experimenting with
> different combinations of keywords. To narrow-down searches further;
> the addition of keywords mentioning specific industries, occupations,
> businesses, people, places, culture, etc. can discern developing
> trends.
> There are many examples of keywords useful for trend-spotting. Five
> examples:
> -- Lately and noticed example.
>  <LONG URL> 
> -- Noticed lately is another variation.
>  <LONG URL> > 
> --'Lately and become' OR 'lately and becoming' can yield
> "important developments" to watch.
>  <LONG URL> > 
> --'People have become' OR 'people are becoming' may infer
> emerging developments from social points of view.
>  <LONG URL> > 
> --'Trend toward' AND 'becoming more' are one of the first
> examples of keywords I developed for trend-watching.
>  <LONG URL> > 
> Entering in keywords which may infer content signaling the emerging
> business & public zeitgeist yields a wide variety search results. In
> quite a few cases, search results don't necessarily have value e.g.,
> individual opinions which don't necessarily indicate emerging trends.
>  In some cases, search results yield specialty websites that are
> fee-only and even restricted access content. Usually, content which may
> be interpreted like "a crystal-ball of sorts" is free.
> Any recommendations for marketing an "inference-like"
> search-engine?
> Thank-you

I like your direction of thoughts, but I think you must go farther than that
in order to make it viable or marketable. I am currently looking into
similar idea and may soon speak to one of the fathers of the semantic Web.
I truly believe it's where we all are headed.


Roy S. Schestowitz      | Useless fact: Women blink twice as much as men
http://Schestowitz.com  |    SuSE Linux    |     PGP-Key: 74572E8E
  3:45pm  up 50 days  3:59,  3 users,  load average: 0.13, 0.12, 0.09
      http://iuron.com - next generation of search paradigms

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index