Nestor Inc. is touting Prism, its new pattern-recognition software, as a breakthrough in credit and debit card fraud prevention.
The Providence, R.I.-based company, which introduced the system at last week's Card Securities Conference in Chicago, said Prism adds to the neural-network technology widely used for fraud detection.
"We've gone beyond just fraud reduction" by adding monitoring systems other than neural networks, said Sushmito Ghosh, Nestor's vice president of risk management.
Neural networks are meant to mimic the learning and pattern-recognition functions that the brain performs. Credit card banks have reduced fraud losses by subjecting transaction data to neural pattern analysis.
Prism uses advanced pattern-recognition technology to read an electronic stream of card transactions and detect accounts or transactions with a high probability of fraud. The software combines neural technology with features such as statistical analysis and rule-based systems of fraud detection to reveal suspect amounts or transactions, the company said.
The rule-based component compares each transaction with a set of conditions that embodies profiles of suspect activity. For example, it monitors how many times a card has been used in one day and whether the card passes its credit limit in that same day.
"Since fraud is a transient behavior, once the fraud-prevention manager detects a fraudulent scheme, it often changes," said Mr. Ghosh. "Prism learns changing trends in fraud and adapts to new fraudulent schemes."
MasterCard, hoping to assert its own presence in this area, in May hired Los Alamos National Laboratory in New Mexico to do pilot tests of systems going beyond the traditional neural networks - systems for MasterCard members only.
HNC Software Inc., a Nestor competitor, has led the commercialization of neural technology for the card industry, working with most of the major card-issuing banks, processors, and Visa U.S.A.
In recent years, HNC, which is based in San Diego, has expanded its focus into merchant risk analysis to help banks and merchant processors ferret out money laundering, skimming, and other improprieties at the store level.
Nestor said it believes Prism will push it ahead of HNC in the credit card fraud detection industry. "We've been competitors for the last five years," said Mr. Ghosh, "and we think we have a much better product."
An HNC official disagreed. "We have an excellent system that allows for real-time detection of fraudulent activity," said Patsy Campbell, director of marketing, "plus we offer a lot more than just neural networks."
Prism operates on mainframes, Unix workstations, or personal computers on a network, Nestor said, and can be integrated to work with all software applications. It can be adjusted to define priority levels and identify fraud by card product, geographical region, and type.
