Tuesday, April 19, 2011

Focus (or lack thereof) on data mining

If you've never been to Oracle OpenWorld, let me assure you that it is huge. But is it too huge? Unless your name is Larry or Safra, the conference probably vastly exceeds your area(s) of interest. If you had to sum up what Oracle OpenWorld is about, you'd probably be reduced to muttering something about "the stack." Repeatedly.

Of course, Oracle OpenWorld isn't the only event that promotes such a broad range of interests. A couple of decades ago, I remember receiving an invitation to a conference - it may have been COMDEX, or it may have been something else. The invitation asked the question "Who should attend this conference?" As you read through the invitation, you discovered that the conference organizers thought that EVERYBODY should attend the conference. Whichever conference it was, it's no longer around, and I'd be willing to bet that part of the problem was that it tried to be everything to everyone, which ended up making it not that compelling a conference for anyone. (Unless, of course, you're into booth babes.)

Side note: the COMDEX 2004...um...postponement was discussed at the time here. While the bursting of the dot-com bubble was involved, a major reason was the withdrawal of key vendors from the exhibit floor. As of today, COMDEX is a virtual conference.

I subscribe to a biometrics mailing list, and someone on the mailing list invited all of us to a conference on data mining.


Well, because.

The European Conference on Data Mining (ECDM’11) is aimed to gather researchers and application developers from a wide range of data mining related areas such as statistics, computational intelligence, pattern recognition, databases and visualization. ECDM’11 is aimed to advance the state of the art in data mining field and its various real world applications. ECDM’11 will provide opportunities for technical collaboration among data mining and machine learning researchers around the globe.

OK, collaboration is a good thing, if I'm collaborating with someone in a fairly related field. Returning to the Oracle OpenWorld example, I attended OOW as a database user. Now I was certainly interested in web servers and operating systems, and maybe a little bit in analytics, but most of the conference was not relevant to me.

So who should attend the data mining conference? Let's look at the topics of interest:

Core Data Mining Topics

- Parallel and distributed data mining algorithms
- Data streams mining
- Graph mining
- Spatial data mining
- Text video, multimedia data mining
- Web mining
- Pre-processing techniques
- Visualization
- Security and information hiding in data mining

Data Mining Applications

- Databases,
- Bioinformatics,
- Biometrics
- Image analysis
- Financial modeling
- Forecasting
- Classification
- Clustering
- Social Networks
- Educational data mining

Now that is a broad array - or an n-dimensional array - of topics. It almost sounds like a joke - "A data miner, a biometrics guy, and a social networker walk into a bar."

However, the conference registration (not counting hotel, travel, etc.) is less than 700 EUR, so perhaps the cost could be justified for some attendees.

If I'm completely off base, and this is something that interests you, go to http://www.datamining-conf.org/. The conference will be held in Rome on 24-26 July.

But from my perspective, even if I were interested in data mining from a biometrics perspective, I'd have to think long and hard to decide if this would be the best thing for me to attend.

Even if there were Italian booth babes.
blog comments powered by Disqus