Data Mining - worth the investment?

[Irishphil]

Hi,
I am currently working in 'data warehousing' for a company specialising in corporate finance and am interested in finding out more about data mining. Currently my team is responsible for the production of monthly standard reports from a transactional system, we also set up reports that the users can do a bit of ad-hoc analysis on. I see developing data mining functionality as a logical extension to the warehouse we have in place, but am not sure of the value it could add to our processes.

Does anyone out there have links to helpful case studies on the subject? I am reasonably familiar with the concepts around the various algorithms used; the sort of thing I'm looking for would be to justify the investment that would be required. Case studies by particular companies in the same field showing what they can now achieve from applying data mining would be ideal!

Thanks,

Phil

 

[Predictor]

I imagine one could find many case studies online, using the usual search engines (Google, Teoma, Dogpile, Yahoo! etc.), but I think the central issue in any of them is classic cost/benefit analysis. Most data mining efforts are predictive modeling projects, so the benefit can be characterized as the cost reduction/sales increase/whatever, given a certain increase in predictive accuracy over the current process. One normally doesn't know exactly what accuracy to expect from the model ahead of time, but one should be able to guess as to what level of accuracy would be of any substantive benefit. This still leaves the hard part of guessing how much accuracy improvement may be realized via data mining.

 

[bkj123]

Phil -

Sounds like a great opportunity to do a Proof of Concept on your own or even better with a vendor.

- Brian