No negatives

[mahound]

Hello,
If I have a data set where only positive cases are labelled as such, and all the other ones are labelled as "Unknown", which approach should I take in order to extract some kind of classification rule?
Every possible strategy that I find implies knowing apriori a classification rule (which is exactly what I want to find).

Any suggestions?

Best regards,

Pedro

[editor]

I suggest you try some clustering algorithms and see if structure emerges.
One-class SVM was also helpful in such situations - google One-class SVM for more.