Researchers detect fraud with highest accuracy to date

Sep 18, 2012
This graphic visualizes how MetaFraud is able to increasingly distinguish between fraudulent (red) and legitimate (blue) firms from one stage of the framework to the next.

(Phys.org)—Researchers from Brigham Young University have helped create the most robust and accurate fraud detection system to date using information from publicly available financial statements.

Using business intelligence software that learns and adapts as it processes data, a team of professors from the Marriott School of Management developed a model that correctly detects fraud with 90 percent accuracy.

"We've improved on 30 years of research in terms of accuracy in capturing fraud patterns," said Jim Hansen, information systems professor and study co-author. "This improved detection is crucial given the broad societal costs of management fraud."

Major fraud scandals at Enron, WorldCom and several other firms around the turn of the century were the catalyst for a 78 percent drop in the between 2000 and 2002.

Despite changes to internal control procedures following those scandals, fraud continues to plague economies throughout the world. In 2008, fraud was discovered when filed for the largest bankruptcy in U.S. history, which dropped the stock market 22 percent and contributed to the most recent recession.

This video is not supported by your browser at this time.
Fraud test finds red flags in company data.

Lead author Ahmed Abassi of the University of Virginia, along with Hansen and BYU information systems colleagues Conan Albrecht and Anthony Vance have spent several years developing their fraud detection tool, "MetaFraud."

The MetaFraud framework is comprised of several base-level "learners" that feed their results into a "meta" or overarching business intelligence algorithm that learns and adapts over time.

Vance likens it to an army general who dispatches spies and informants to carry out reconnaissance and then report their findings back to him. These digital spies report back on patterns observed in annual data, quarterly data, organizational context and industry context so the "general" can generate an assessment of whether or not a firm is fraudulent.

"When people commit fraud, it invariably hits the numbers," Albrecht said.

The researchers funneled 9,000 instances (instance = financial information for one firm for one year) from more than 15 years through MetaFraud, and compared its predictions of fraud with known fraudulent firms identified by the SEC. MetaFraud correctly predicts fraud with 80 percent accuracy, and over 90 percent accuracy when MetaFraud reports a high level of confidence, which it was able to do 70 percent of the time.

"This is a very low-cost method of taking data that's already out there, already public, and then in a few minutes calculating, with a fairly high degree of certainty, if fraud is happening in that company," Vance said.

The details of the algorithm and computer models appear in an article published online ahead of print in a special business intelligence research issue of MIS Quarterly, the premier journal.

In making their MetaFraud algorithm available, the BYU researchers hope to equip investors, external auditors and regulators with a tool that accurately assesses a company's risks.

"We've got to be able to stop and corruption – or at least discover it," Albrecht said. "There are people all over the world working on this and there have been a lot of papers published on this. We haven't solved the problem yet, but this is an important breakthrough."

Explore further: Modeling the ripples of health care information

Related Stories

Cybercrime against businesses 'explodes'

Nov 29, 2011

More than one in three companies say that in 2011 they have been victims of fraud, a report said on Tuesday, with reports of cybercrime targeting businesses around the world skyrocketing.

Romanians charged in US with computer fraud

Dec 09, 2011

Four Romanians have been charged with hacking into the computer systems of hundreds of US merchants and making millions of dollars in purchases with stolen credit card data, the Justice Department said Thursday.

US charges 91 in wave of health fraud cases

Sep 08, 2011

US authorities have charged 91 people over some $295 million in alleged fraud schemes related to Medicare, the government-run health program for seniors, the Justice Department said.

Recommended for you

Forging a photo is easy, but how do you spot a fake?

Nov 21, 2014

Faking photographs is not a new phenomenon. The Cottingley Fairies seemed convincing to some in 1917, just as the images recently broadcast on Russian television, purporting to be satellite images showin ...

Algorithm, not live committee, performs author ranking

Nov 21, 2014

Thousands of authors' works enter the public domain each year, but only a small number of them end up being widely available. So how to choose the ones taking center-stage? And how well can a machine-learning ...

User comments : 13

Adjust slider to filter visible comments by rank

Display comments: newest first

tadchem
5 / 5 (3) Sep 18, 2012
Would this be applicable to government agencies?
jscroft
2.3 / 5 (3) Sep 18, 2012
Shhhhhh!!!
Jeddy_Mctedder
2 / 5 (4) Sep 18, 2012
The second the algorithm becomes transparent it becomes a tool for cheaters to help confuse people. The reality is that the myrh that fraud is hard to detect is a big lie. We know where the fraud was. We bailed it out. Put no one in jail.

If i had artificial intelligence algorithms to detect fraud id use it to discover which prosecutors and politicians are bought off. The truth is almost all of them are.

Fraud in the markets is not only systemic but it all traces back to the zero interest policies of the tbtf owned federeal reserve. They decide which frauds get zero interest loans to paper over. And which frauds get tucked away like mf global.
Meyer
not rated yet Sep 18, 2012
Out of curiosity, for which companies does it report a "false" positive with high confidence? Please reply or forward the list directly to the SEC, FTC, FBI, TSA or whoever deals with this stuff. :)
antialias_physorg
5 / 5 (1) Sep 18, 2012
The second the algorithm becomes transparent it becomes a tool for cheaters to help confuse people

Since it's a learning system there's no specific algorithm behind it ('only' a learned set of weights applied to input data)
Caliban
1 / 5 (1) Sep 19, 2012
The second the algorithm becomes transparent it becomes a tool for cheaters to help confuse people

Since it's a learning system there's no specific algorithm behind it ('only' a learned set of weights applied to input data)


And, just like any tool --it is only useful if it is put to use.

Why do I have very little confidence that this tool will be adopted and put into use by, say, the IRS, FBI, SEC, or even the Justice Department?

But this is a sad, sad day for any one in an organization that is employed below the VP level...

rsklyar
3 / 5 (4) Sep 19, 2012
It would be notably useful "to stop fraud and corruption – or at least discover it" also among some US professors: https://connect.i...sr/blogs
antialias_physorg
not rated yet Sep 19, 2012
Why do I have very little confidence that this tool will be adopted and put into use by, say, the IRS, FBI, SEC, or even the Justice Department?

But you can bet that it will be put to use by insurance companies and investment bankers.
So while there may not be legal recriminations the result will be the same: fraudulent companies will go bust (hopefully) sooner than later.

But this is a sad, sad day for any one in an organization that is employed below the VP level...

If you're employed in an organization that has fraudluent VPs/CEOs then you're shafted either way. Not much you can do about it (and precious little chance that the top echolon will get billed for their fraud).

It may not change much in the current state of affairs, but it may prevent future ocurences (or at least detect them earlier so that less damage is caused). If it does that then it's already worthwhile.
Caliban
5 / 5 (1) Sep 19, 2012

If you're employed in an organization that has fraudluent VPs/CEOs then you're shafted either way. Not much you can do about it (and precious little chance that the top echolon will get billed for their fraud).

It may not change much in the current state of affairs....


AA,

This is exactly my point. The fraud that really matters is practised top-down, from the very highest levels of the banking/finance/insurance industries --which has just recently been demonstrated to us, most noticeably beginning in September of 2008.

The occasional fraud in business that occurs below the VP level is so miniscule in extent as to be nonexistent in comparison to the fraud perpetrated at the command-collude-and-control level.

And just for further clarity --fraud practiced on the petty level is of course still fraud, but it is a pervasive evil that the body capital can and does live with, as evidenced by the generally robust global economy over the last 60 or so years.

antialias_physorg
not rated yet Sep 19, 2012
Maybe there's still an out here: What do people do who engage in fraud? They reap lots of money. So far, so bad.

BUT these people are also aware that fraud happens (naturally). So they are interested in making sure that they don't reinvest money they just grabbed into others' fraudulent enterprises.

It might be that fraudsters are the ideal group to want to use this method against other fraudsters...and eventually they just starve each other of funds.

Wouldn't it be hilarious if that happened?
Alburton
not rated yet Sep 23, 2012
MetaFraud is such a bad name for a fraud detecting system.Doesn't it quite literaly indicate that the program itself is a fraud based on detecting frauds,making it a new layer of fraud?

kochevnik
not rated yet Sep 23, 2012
BYU should stick to things they're good at, like getting 20 year olds married, spying on Americans at the NSA and persecuting gays.
julianpenrod
1 / 5 (2) Sep 24, 2012
Thre are a number of points to notice.
Among other things, the software claims to be able to spot fraud with at least 80% accuracy. The "proof" is questionable. The way it's phrased, it seems to suggest that 9000 firm years of data were fed through MetaFraud, whether 15 years worth for 600 firms or 1 years worth for 9000 firms is not specified. But it is acknowledged that predictions of fraud for those firms correctly identified known fraudulent firms at least 80% of the time. But, these years that fraud is predicted, are they years that the SEC asserts those firms were acting fraudulently? What about the 20% that weren't recognized as fraudulent? How did they manage to hide it?
And what of the 20% of predicted fraudulent firms that, supposedly, haven't been brought up on charges?
Are they, perhaps, really fraudulent?
Shouldn't people be warned of their identities?
Maybe MetaFraud was right in 100% of predictions, some just haven't been charged yet!

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.