Coming Attraction

Whether you're a serious investor or just dabble in the stock market, you might want to take notice of 16-year-old Ben Brinkopf. The Canterbury High School sophomore took home Best of Fair honors at the Thomas A. Edison Science and Engineering Fair for his project, Integrating Neural Networks with Pattern Recognition to Predict Stock Market Movements. In plain English, that means Brinkopf created a computer network to forecast the momentum of individual stocks.

"Economists have traditionally used linear systems to determine momentum-if this happens, that will result," he explains. "Neural networks take large amounts of information and find several patterns where linear algorithms only find one." Brinkopf compared the price-and, in some cases, the volume and indices-of 12 different stocks. "His project may help people determine where to invest in the stock market," says Canterbury research adviser Betsy Glass.

What were his initial findings? "In most cases, I predicted the rise and fall of the stock," says Brinkopf. "My results indicated that using a variety of factors is significantly more accurate than just one."

While there are a number of books published on this subject, Brinkopf concedes there's still research to be done before his theories are applied in the real world. "Stock information needs to be real time, not delayed. And neural networks don't predict people's reaction to news stories," he says. "What might be someone's bad news is someone else's good news."

So how did this cross-country runner, sophomore class president and aspiring pilot become interested in combining neural networks and the stock market? "I've always been fascinated with both-especially the stock market," he says."

As for whether or not his research will become commonplace in the investment industry, "time will tell," Brinkopf says.