What is a Topic?

The most mysterious term that I encountered a lot recently is topic. I have no idea how to define it and, neither seem the weblog research proposals that suggest finding the topic of a post is something worth doing. Being on holiday currently, and given it was raining and snowing outside, I tried to apply the notion of "topic finding'' to weblog conversations (see also: here, and here).

In Making a Difference terms used by bloggers in a "community" were compared with the objective to determine terms (topic?) that interested individual bloggers but not the entire community. This time, the source was a weblog conversation (loosely a collection of linked posts in a community) and determine what distinguishes a conversation compared to what the community usually blogs about (topic?). Three examples are shown in the table below (first number is frequency of term in the conversation, second number is score of term with respect to conversation vs. other posts inside the community).

Conversation 1	Conversation 2	Conversation 3
social tool [33, 19.3] tool [89, 18.5] blogwalk [40, 15.8] conference [54, 14.0] people [117, 13.6] blog [127, 12.6] blogtalk [34, 12.3] symposium [18, 11.9] experience [43, 11.8] event [33, 11.0]	sigmund [36, 31.9] blog [183, 26.7] blog research [29, 20.6] research [63, 20.1] analysis [37, 20.1] post [79, 19.5] conceptualisations [21, 19.0] paper [39, 16.2] rss [35, 12.9] study [30, 12.0]	presence [90, 38.3] skype [84, 32.8] communication [59, 27.5] im [40, 26.0] communication tool [23, 21.4] blog [129, 19.5] medium [31, 19.0] social [70, 18.0] tool [58, 17.4] quality [34, 15.7]

Conversation 1 is about the BlogWalk 3.0 event (experience, conference, symposium) which took place just before BlogTalk.

Conversation 2 is about Sigmund (the Sigmund paper is called "Shared Conceptualisations in Weblogs").

Conversation 3 is about Skype.

Although I'm still at a loss explaining why the methods I use work so well, it is clear that the notion of what a "topic" is has not been solved (is it the conjunction of the "most significant terms against some baseline"?). I'll check with my colleagues once I've returned from holidays :-).

One thing is crystal clear though: weblog research is more interesting when one concentrates on a well-defined subset of the blogosphere.

Comments

[Lilia]

Anjo,
wondering if it makes sense to use ratio of the frequency of term in the conversation / number of posts in a conversation? Wouldn't change the results for a specific conversation, but would make comparison easier...

Also - I'd expect that "topic" should be well represented with Sigmind picture for the given terms.

[Aleks]

My own definition is:
==
A topic is a semantic annotation of a document that people instinctively use to search and organize information. While we can semantically annotate many facets of a document, the notion of topic covers the document as a whole: it is the conceptualization of the document with the greatest breadth. When we seek information, our primary intention is usually to discover information on the same topic as is the topic of our problem. Topic can be distinguished from other descriptors of a document, such as language, style, punctuation, grammar, reputation, graphical design, author, date, and so on.
==

BTW, you can see how topics can be visualized graphically at [cosco.hiit.fi]. This is based on the open-source software for identification of topics from [componentanalysis.org].

[Anjo]

Lilia, I'm using the ratio. Stupid I missed the Sigmund metric on first hand (perhaps I am on holiday :-)).

Aleks, I think your definition is sort of the right direction, but fail to see how to operationalise it in any rigorous manner. Browsing the links you provided I became a little wiser :-). Thanks.

[JoseAngel]

Perhaps one cannot dissociate the definition of what is a "topic" from what does "going off-topic" mean in a specific context. Topics can grow, evolve, as further aspects are unearthed or perceived, and they can be contested... that is, you cannot take the topic as a given, it is shaped by the ongoing conversation.