Datamining for outliers DOES NOT WORK
Well if the US can stick their neb in, so can I. Look, that idiot in uniform, General Alexander, created a system of data mining for terrorism. Terrorists are outliers and data mining for outliers NEVER WORKS.
And filtering based on this faulty idea, means you're not targetting 'radical muslims' or whatever, you're targetting *everyone*.
Your filter looks something like Everyones_Data.FilterByReligion(Muslim|Jedi).FilterByComments(Mumsnet|elReg|Slashdot)
You can't detect religion, you filter by 'checkbox on the electoral form' or 'gps location next to mosque'.
You can't detect radicalism, you substitute 'comments on forums' or some such garbage.
You can't detect 'acts' so you detect 'speech' instead, people reading and writing stuff on the internet.
i.e. you are attacking freedom of association, and freedom of speech, and you're bypassing the judicial process, eliminating right to judicial process.
Would you lot just fnck off. Really you're not the defender of your country, you're the underminer of your people. The next step in this garbage is to define the output of the filter as the crime, i.e. terrorism becomes 'said things that might aid terrorists'. i.e. speech as terror!! All to cover the loss of face from investing in a process so flawed, and to protect your big data budgets.
Did your algorithm spot Florida/Gay terrorist? Of course not, he's an outlier. Next time your algorithm will select lots of people who happen to live near gays to target for false prosecution.
Data mining for outliers DOES NOT WORK. Austria lawyer is right, you are wrong. Obey the basic rights because they have stood the test of time, and your shit wouldn't survive even a cursory analysis without the 'national secrets' you hide behind.
Biting the hand that feeds IT
