As I guess most people are aware, there are constant discussions on how strict the database should be and how much "common sense" should (or can) be applied etc.

But with the following data available:

The "strict by the rules" master database

Uploaded collections with users local data

it might be possible for a standard database mining algorithm (as build into the database, not something you have to program) to calculate how likely it is I would want another value than the entry in the master database in a specific field. It might be more tricky with ordered "list data" like cast etc, but I would not be surprised if that could be done as well, the data mining algorithms appear to be pretty smart.

At least it would take away the discussion if something is "common sense" or not - because if it is "common" the data mining will show it.

The only thing useful for data mining is for invading privacy....err, I mean searching for terrorists.

And before anyone flames me for the comment, I'm just kidding. It's just a joke. I think this is a pretty good idea. 

Karsten,


It looks like you are thinking about check of which value is used by most people which relates to datamining algoritms the same way as counting on your fingers relates to super computers.

Even if it was a "finger counting" algorithm it might work with the proper constant, which would obviously not be 50% but a significently lower number. No, I do not know what it would be, that is something that has to be testet, and most likely you would have to be able to specify some preferences to indicate if you favour "common sense" over "accuracy". Obviously the percentage of people "suppressing" a "common sense" value in their database could be taken into account as well.

I suspect it might be problematic to run this on profiles very few people have uploaded, but then it should just be disabled until a sufficient amount has been reached.