Friday 23 October 2009

Knowledge Discovery in Databases: what does exactly lie in our Databases?

In this brief text, Peggy Wright gives an overlook of Knowledge Discovery in Databases (KDD) and of technique to draw knowledge out of a large amount of data.

The text is available here.

KDD, as data mining, may be tools and as such ethically neutral. However, how should the "data pre-process", which implies removing "meaningless data" be understood? Indeed, it implies that some behaviours, pathologies etc. can be neglected. While it may be accepted if you think of KDD as a search for the answer to a specific question, it sounds pretty shocking to my ears if you think that Data-Mining aims at finding "patterns" and regularities.

Actually, pre-processing data entails the removal of some patterns and as such the selection, I would even go further by saying the "standardisation", of a certain type of behaviour.
Therefore, I think that trying to unfold the ethical stakes of KDD, even though knowledge discovery is a natural tendency, is crucial not to turn our society into what G. Orwell called "Big Brother".
Such a reflection is available here.

No comments: