Folksonomy to Ontology

So after presenting all my views on why it should be possible to transform tag clouds to structured data entities (ontologies of sorts), I have gone about devising a method to do so.

When it is ready, I will release a Protege project showing how it is done. In the meantime look at my Protege page on this blog, which is where the project will eventually go.

Advertisements

5 comments on “Folksonomy to Ontology

  1. Greg says:

    Hi Csaba,

    Just to back it up a step, have you seen tagthe.net? They take a HTML document and extract (with varying levels of success) the tags therein. They’re looking for add-ons and extensions, and I wonder if it might be useful for building tagsets?

  2. Csaba says:

    Hi Greg,

    Thanks for the pointer. A very interesting site. Have to say, on first impressions, VERY BIZARRE!!

    I mean, the whole idea of tags is that they are user generated and reflect the perspective of the community, rather than some sort of keyword extraction.

    Or am I missing something somewhere?

  3. Greg says:

    It does seem a little odd. I think the idea is that the logistics of tagging stuff means that there’ll always be more things that can be tagged than people willing to tag them. To address the imbalance, perhaps some half-arsed (or even quarter-arsed) tagging is better than none?

    This leads to an interesting empirical question: in the tagging marketplace (where attention is scarce), will a “badly” tagged object get tagged sooner than an equivalent untagged one? If the answer is yes, then maybe such services have a role in goading (human) taggers into action?

    My experience with wikis suggests the answer is yes. People seem to be irked more by misinformation than its absence.

  4. Steven Forth says:

    It may well be productive to have mixed tagging straegies in the same collection of tags. I see nothing wrong with mixing human tags with a variety of system generated tags. This could even have some interesting emergent properties.

    I look forward to seeing your approach applied using Protege. Deriving ontologies from social tagging is likely to be a very productive area. But then I can use the derived ontology to help organize my own tagging, can I not?

  5. Csaba says:

    Sorry I have been neglecting my blog .. and my work to some extent. I have moved jobs and countries (continents, acyually!). Hopefully I will be back in action soon.

    But I think I have to disagree with you. The whole reason I became interested in folksonomy (wearing my cogntive psychologist hat here) is that they reflect un conscious (in large part) human processes involved in comprehension and classification. To me, they are an un contaminated view into the way human minds organize their environment.

    As a computer scientist now, I think exploiting these observations can have great benefit for facilitationg search and discovery. But the last thing I would want to do is interfere with the process generating the tags in the first place.

    User interface issues might already cause some issues .. but system generated tags seems to me to fly directly in the face of the whole enterprize.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s