It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Why Data Mining is Useless Against Terrorists

page: 1
0
<<   2 >>

log in

join
share:

posted on May, 24 2006 @ 10:18 PM
link   
There's been a lot of fuss about data mining lately. The NSA listens to everyone's 'phone calls, reads all emails... and it's all to protect us from those nasty terrorists.

Well, you might think that was all BS... and if you do, this article* is the one for you. In it, Floyd Rudman goes through some nifty statistical analysis to demonstrate that, not only is data mining an absolutely useless tool for identifying terrorists, but that the NSA must know that that is the case.


The whole NSA domestic spying program will seem to work well, will seem logical and possible, if you are paranoid. Instead of presuming there are 1,000 terrorists in the USA, presume there are 1 million terrorists. Americans have gone paranoid before, for example, during the McCarthyism era of the 1950s. Imagining a million terrorists in America puts the base-rate at .00333, and now the probability that a person is a terrorist given that NSA's system identifies them is p=.99, which is near certainty. But only if you are paranoid. If NSA's surveillance requires a presumption of a million terrorists, and if in fact there are only 100 or only 10, then a lot of innocent people are going to be misidentified and confidently mislabeled as terrorists.

The ratio of real terrorists to innocent people in the prison camps of Guantanamo, Abu Ghraib, and Kandahar shows that the US is paranoid and is not bothered by mistaken identifications of innocent people. The ratio of real terrorists to innocent people on Bush's no-fly lists shows that the Bush administration is not bothered by mistaken identifications of innocent Americans.

Also, mass surveillance of the entire population is logically plausible if NSA's domestic spying is not looking for terrorists, but looking for something else, something that is not so rare as terrorists. For example, the May 19 Fox News opinion poll of 900 registered voters found that 30% dislike the Bush administration so much they want him impeached. If NSA were monitoring email and phone calls to identify pro-impeachment people, and if the accuracy rate were .90 and the error rate were .01, then the probability that people are pro-impeachment given that NSA surveillance system identified them as such, would be p=.98, which is coming close to certainty (p_1.00). Mass surveillance by NSA of all Americans' phone calls and emails would be very effective for domestic political intelligence.


Now that article's what I call denying ignorance.

* The article is, as I post, on the front page of the website. It will be moved and you should look in the LH sidebar for Floyd Rudman. I'm not sure how they'll title the article. If I have to change the link, I will, but it might not be for a day or so.



posted on May, 27 2006 @ 05:38 AM
link   
I can't edit my original post so I'm reposting the article (which has boon moved) in this link.

It would be nice to know that someone has seen this... I think it's an important perspective on the subject.



posted on May, 27 2006 @ 05:55 AM
link   
They're trying to find out how much we know about their sleezy operations.

How about the Billions they have stashed away in Swiss bank accounts.

How about the drug runs to finance their campaign.

Election fraud, lies and deception, supression of technology all of it.

They are trying to shut down leaks and trace sources and find out who's organizing
against them.

I'll bet you they have traces of the home cell and office phones of every congressman
and their staff or anyone powerful enough to organize articles of impeachment.



posted on May, 27 2006 @ 06:18 AM
link   
that piece by Floyd Rudman as you cited...
is basically a 'mirror image' of the spin justifying Intels' position...
both views are based on the "Baffle Them With B.S." model...

any accountant knows that its how you present/spin the data as 'facts'

the court of public opinion is what their both trying to convince, sway.

(and) At this early stage, i'm leaning towards the Intel community side=
let them develop expanding rings of connections
let them construct 'lattice frameworks' of persons capable of societal harm
let them deduce that Six Degrees of Separation is a truism



posted on May, 27 2006 @ 06:30 AM
link   
Of course they know that's the case, the terrorists are an excuse for the authorities to spy on normal people.

One that protest against the government, peace activists etc People who have done nothing wrong.



posted on May, 27 2006 @ 08:23 AM
link   
I seem to remember that the problem pre9-11 wasn't not having dots but rather not connecting those dots.
To paraphrase, we have met the terrorists, and they are us.



posted on May, 28 2006 @ 04:24 AM
link   
I suspect that that is true, but perhaps not in the way that you mean.



posted on May, 28 2006 @ 08:02 AM
link   
I think the piece that started this thread is basically an op/ed piece, devoid of any actual facts.

"Datamining" is only relevant when you already have a profile (which the NSA does have. The real questions ought to be 1) How good is the profile we're using, and 2) what "accessory facts" are useful for locating a persons who is about to but has not yet put a terrorist plot into motion.

Here's a list of items I'd go for:

-multiple travels to nations of interest: Palestine, Iran, Iraq, Afghanistan, Pakistan.

-cell phone or email contact with embassies or diplomatic personel/embassy workers from states known to harbor terrorists

-lack of any visible means of support (job)

-people who buy weapons or explosives, or their precursors, and fit some other category on this list

-foreign national who is a college student, with lack of academic progress toward a degree. (several of the 9-11 hijackers fit this)

-travel to a city where the person has no known connections: no history of prior phonecalls (friends/relatives, business) (9-11 jackers fit this too)

-purchase of an inordinate amount of video equipment, or is seen filming a security location, or security personell. (suspected terrorists in Dallas TX caught filming the downtown skyline from various angles)

-travel to a distant American city, with expenditures for food and lodging, but no relatives or apparent goal. (i.e., suspect lives in Miami and travels to Seattle. Stays in a downtown hotel, and uses credit cards for food and a rental car. But makes no cell phone or hotel room phone calls, doesn't go to movies or "do" anything else that takes money while in town.)

-receives money, or sends money, out of the country.

- is a roomate or frequent phone or email contact of someone who fits most of the above categories.

And that's just the list I came up with over coffee. Not that those people are guilty. But they should get a free interview with the govt, just to make sure everything's OK.

Do you think the government shouldn't be looking for people who fit more than 1 of the above categories??? Shouldn't they at least consider that such people are terrorist?

If you wouldn't be profiling from MY list, maybe you can come up with a better one? One that weeds out the innocent?

Or maybe nothing the government does will ever be good enough for you.

.



posted on May, 28 2006 @ 11:31 AM
link   
What's the difference between "op-ed" and "analysis"?

Certainly opinions are expressed in this piece but to say there are no facts is nonsense. There are facts, and they are put into a coherent context to advance a specific argument, namely, that the tool of datamining is useless for finding terrorists, but good for domestic surveillance of dissent. I have no intention of duplicating a perfectly cogent thesis here that is already admirably stated in the article.

I'd say he does the job, and it's an argument you don't address. To say that "there are no facts" is to make me suspect that you didn't read the article, or that if you did, you didn't follow his argument.

How do you know the NSA has a profile? Do you know what that profile is? Do you work for the NSA? The more I read your post, the more I think the actual point of the argument - that profiles are irrelevant because statistically that kind of sweep is guaranteed to give you high rates of misidentification - has passed you by.

Are you aware that government agencies like the FBI or NSA have been targeting peace campaigners and environmentalists? Your government is moving closer and closer to high-tech totalitarianism and you seem utterly happy for that to happen. The author makes the telling point that these approaches are indicators of paranoia rather than useful approaches to finding terrorists, and that the government cares little that innocent people are, at a minimum, put on no-fly lists or at worst, picked up and sent to another country for torture (see this thread for the relevant story).



posted on May, 30 2006 @ 08:05 PM
link   

Originally posted by rich23

that profiles are irrelevant because statistically that kind of sweep is guaranteed to give you high rates of misidentification - has passed you by.



First, I do understand that the op/ed piece is claiming that "that kind of sweep is guaranteed to give you high rates of misidentification." I just don't agree with it; I also don't believe the author offered any really persuasive argurments for such a conclusion.

First, why is it "guaranteed" to misidentify people? If the profile doesn't work, it can and should be tweaked to provide better results. Like most sorting mechanisms, any profiling system is only as good as the paramaters you apply.

Racial profiling doesn't work, but that's because race isn't an accurate predictor of a crime. On the other hand, some of the identifiers I suggested are excellent for picking out the sorts of terrorists that DHS has already nabbed, since I was thinking of current arrestees when I came up with my post above.

I would argue that datamining in general, and applying the results to profiling, works incredibly well. If you don't believe me, then just take a college class in advertizing, and read the textbook. Advertizing companies developed datamining, precisely to predict future individual behavior. And they haven't gone out of business as of this evening. So, if you (or the author of the editorial) want to claim datamining/profiling are useless, fine. But don't be surprised when people who understand the process are sceptical of such assertions.



Originally posted by rich23

Are you aware that government agencies like the FBI or NSA have been targeting peace campaigners and environmentalists?



Yes, I am.

Are you aware that some "peace campaigners" have received money from office holders in middle eastern governments?

$777,000 from Saudi Arabia via Egypt

Foreign ties and political aims of various US-based "arab civil rights" associations.

and that's just the first page of a google search of "peace activist + donation + foreign." A person who was interested might turn up all sorts of interesting information on the "peace" movement.

I'd be fascinated to hear from you about what system you'd use to find terrorists. Or maybe you don't believe there are any.

Do you think America has actual enemies? I mean other than its own government?



posted on May, 30 2006 @ 08:09 PM
link   



The ratio of real terrorists to innocent people in the prison camps of Guantanamo, Abu Ghraib, and Kandahar shows that the US is paranoid and is not bothered by mistaken identifications of innocent people. The ratio of real terrorists to innocent people on Bush's no-fly lists shows that the Bush administration is not bothered by mistaken identifications of innocent Americans.



How does the author know what the ratio of innocent people is?

Doe HE work for the NSA? Has he interviewed ALL of the Gitmo detainees?

Has he interviewed EVERY PERSON whose name is on a "no fly list?"

If the answer is "no," then he's talking out of the wrong end.

.



posted on May, 30 2006 @ 08:51 PM
link   
I really don't think datamining is the crux of the dilemma with this spy-program. The problem boils down to a lack of oversight on operations being run by the NSA. Who will step in when the NSA, or the government in general, starts trying to covertly crack-down on dissidents (which it already has and is probably still doing)? Apparently, warrants are too cumbersome and the only way to defeat the "terrorists" is to let an ultra-secretive agency run amuk. Great plan!



posted on May, 31 2006 @ 01:20 AM
link   
Well .....data mining is absolutely useless in tracking terrorists since it refers to programs hidden in others that monitor your online activity and reports it to businesses, as to what certain groups are most likely to buy.

so...since data mining programs are hidden in programs like p2p networks etc. sont make much sense to me.

a Trojan that allowed you to access computers would be far more valuable tool or simple sniffing out the data packets over hundreds of networks would be too.



posted on May, 31 2006 @ 07:36 AM
link   

Originally posted by XphilesPhan
Well .....data mining is absolutely useless in tracking terrorists since it refers to programs hidden in others that monitor your online activity and reports it to businesses, as to what certain groups are most likely to buy.



Actually, that is only one (trivial) application of datamining.

Wikipedia's definition is



"The science of extracting useful information from large data sets or databases"


Examples would be:

1. The supermarket offers you a "saver tag" to put on your keychain. Every time it is scanned with your purchases, you get a 5% discount. It also groups all of your purchases together, and develops a profile of your spending habits. Then they advertize to you based on your individual tastes. They also know to put the doritos next to the cokes, since most people buy both if they buy either one. The albertson's chain once sent my wife a pregnancy magazine, after she quit filling her birth control perscription, and buying hygiene products, and we bought four pregnancy tests in one morning.

2. A car dealership purchases a "composite statistical database" of the consumers in the surrounding zipcodes, from the US census. It then buys a database of home values from the local taxing authorities. Next, it merges those two databases with the DMV's registry of automobiles and home addresses. From this information, it can calculate which Census responses came from which homes. (this is a true example. It's in the book "The Naked Consumer." I'm not making it up.) Then, the dealership can calculate your household earnings, and decide whether you could afford a slightly more expensive car. Then you begin to get ads in the mail for the type of car you could afford. The automobile industry has found this works incredibly well. As of about 3 years ago, if you get "car junkmail," it is because you've been specifically datamined.

If THOSE nefarious uses of datamining are legal, why is it so horrible to imagine law enforcement using the same types of information to develop a list of possible suspect.

It's not like anyone who pops up from such a search is put in jail. They are just the subject further investigation.

Here's another one, datamining for possible terrorists.

Mine the lists of people who pay for rental trucks, but who don't have a change of address within two weeks afterward; or who never connect utilities in the city where they've delivered the truck. The list would need to be further culled to remove all the kids going off to college, as well as small businesses making a one-time shippment of production equipment. (I've done both of those myself).

Datamining that information would have caught Timothy McVeigh, as well as the 1993 bombers of the World Trade Center.

So, datamining is far from "useless." It isn't an asnwer, either. It's simply one more tool to predict the odds of someone's future behavior.



posted on May, 31 2006 @ 07:50 AM
link   
First up, this article is just bogus spin. Trust me when I tell you data can be represented to say anything you want, without actually lying or making stuff up. It’s what I do for a living after all.

Secondly, the NSA is not listening to each and every call or reading every email, they don’t have the resources to do that (they have the ability, but not the resources, HUGE difference). They are looking for communication patterns that may help them pinpoint where they need to read and listen too. It’s just good policing, and it’s nothing more than a modern day stake out.

What do cops do? They look for suspicious patterns and behaviors to help them pick out criminals from non criminals.

I would be terrified if they did not look for patterns and trends in all the communications in the USA.



posted on Jun, 2 2006 @ 05:40 AM
link   

Originally posted by skippytjc
First up, this article is just bogus spin. Trust me when I tell you data can be represented to say anything you want, without actually lying or making stuff up. It’s what I do for a living after all.


YAAAYYY!!! Skippy 'fesses up!

Not exactly a surprise though. Every one of your posts I've seen shows you desperately trying to spin a negative into a positive. And I particularly like the way you ask us to trust you while admitting that what you do for a living renders you utterly untrustworthy. I REALLY like the logic of that!

However, just to call it 'spin' is not enough. If you can produce a convincing counter-argument to the one set forth in the article, I'll be glad to listen. But just crying 'spin!' and telling us to trust you is pretty lame, I'm sorry.



I would be terrified if they did not look for patterns and trends in all the communications in the USA.


Well, now, this is entirly covered in the argument of the original article. You remember the part where the guy talks about people being paranoid? There you go. The thinking behind this surveillance, behind torture, behind abduction (sorry, "extraordinary rendition", ha!) is paranoid in the extreme. There's a red under every bed and an Islamofascist in every kebab shop. It just doesn't wash.

The more I see of debates like this the more I'm reluctantly drawn to agree with people like Reich in his Mass Psychology of Fascism. Broadly put, his argument is that many people desperately want a strong leader to take care of them and are so in need of this that they'll give up any freedom, surrender any moral high ground (though spin their way into thinking they haven't), and project their fantasies of the strong leader even on to an utter cretin like Bush.

This is a man who can barely string a sentence together, works barely at all, yet is allegedly in charge of the world's sole superpower. Even the most cursory glance reveals him as a moronic psychopath who gloats over others' misfortune. Yet the Bushies big him up at every turn. One of the weirdest things I've ever seen is noticing how suddenly, after he got elected President, it wasn't OK any more to take the mickey out of him for being dumb. He went from being a moron to being a "Dear Leader" type figure.



posted on Jun, 2 2006 @ 05:52 AM
link   
Surely all the domestic spying programs show a mistrust of the people these agencies and the people they answer to (the government) are supposed to actually protect.
Is the government really so scared of the people that it has to keep tabs on them to root out dissenters?
I'm sure many real terrorists realise this and quit using hi-tech comms a long time ago, relying more on a simple letter or word of mouth instead of an email string, messenger chat or landline/cell call.

Joe stalin would be proud



posted on Jun, 2 2006 @ 06:04 AM
link   
COINTELPRO goes back at least to the sixties, and USG agencies routinely infiltrate peace organisations and lately environmental groups.

The fact that most of the wingnuts refuse to recognise is that their government and media are run by a handful of big corporations who want the playing field tilted heavily in their favour. They will therefore do their best to ensure that activist groups of many different kinds are targeted by the various policing and surveillance agencies.

On a related note, it's known that ECHELON, the surveillance program that keeps track of emails and telephone calls across the rest of the world, has been used for simple industrial espionage on behalf of large US corporations. It's not much of a step to see that government agencies regard any grass-roots organisation as a potential threat to one or another of the lobbying groups.

One of my favourite terms from this area of discourse, btw, is "astroturf", which refers to a fake grass roots organisation created by a corporation or alliance of same. The Wise Use lobby is an example of this and is basically funded by logging companies.



posted on Jun, 2 2006 @ 07:54 AM
link   

Originally posted by rich23

This is a man who can barely string a sentence together, works barely at all, yet is allegedly in charge of the world's sole superpower. Even the most cursory glance reveals him as a moronic psychopath who gloats over others' misfortune.



Bachelor's in History from Yale.
MBA from Harvard

How about you?

And I'd be careful with the phrase "gloats over other's misfortune." I've seen some pretty snide comments emanating from your keyboard. Hardly makes you an impartial judge.

Oh, yeah, you never did respond with the exact ratio of innocent detainees at Gitmo or Abu Ghraib. But hey, why let a dearth of data get between your rhetoric and a paranoid screed, right?

.



posted on Jun, 2 2006 @ 08:13 AM
link   
We have not been attacked since 9/11/01. So whatever the government is doing, keep it up, it seems to be working!

Battlefield intelligence should not be halted.




top topics



 
0
<<   2 >>

log in

join