AI & Neural Networks to Analyze and Select Stocks

AI & Neural Networks to Analyze and Select Stocks2017-06-23T13:51:51+00:00

AI & Neural Networks to Analyze and Select Stocks

Company News and Press Releases

Making The Big Leagues American Century Fund Uses Artificial Intelligence Newton Fund to Bring Neural Network into Realm of Larger Money Managers

By Fred Williams

KANSAS CITY, Mo. – A new mutual fund from American Century Investments is about to provide a major boost to the use of artificial intelligence and neural systems in analyzing and selecting stocks.

American Century officials declined to comment on the offering – called the Newton Fund, citing the “quiet period” following its Securities and Exchange Commission filing. According to the filing documents, the fund will be managed using a “sophisticated system, engineered through advanced signal processing and neural network technologies that attempt to capture and interpret investor behavior toward a company in an efficient manner.”

Artificial intelligence involves the use of “smart computers” that analyze multiple variables and make decisions via reasoning and learning that are similar to human thought. Proponents deny that AI is a quantitative process or “black box” system used to automate a predefined set of investment rules.

Only a handful of investment managers use neural systems, a more advanced form of AI that involves learning and recall.

Proprietary Data

Newton’s AI system, which was developed internally at American Century, uses proprietary data to select stocks of companies “regardless of size, industry type or geographic location.” Artificial intelligence will be used to conduct “technical analysis of investment opportunities” and to “identify companies whose share price patterns suggest that they are likely to either rise or fall in price.” A spokeswoman at American Century declined comment when asked if the system would attempt to mimic American Century’s bottom-up stock selection process. She said the fund is tentatively scheduled for launch later this year.

Although Newton will use neural networks to select stocks, James E. Stowers III, American Century co-chairman and chief investment officer, is listed as fund manager, along with John Small Jr., who has served as an analyst and portfolio manager for the American Century Ultra Fund and is now co-manager of the American Century Veedot Fund. Mr. Small also has a master’s degree in laser-optics physics from the Air Force Institute of Technology, Wright-Patterson Air Force Base, Ohio.

The fact that a heavyweight like American Century plans to utilize AI in managing stocks is encouraging to smaller investment firms already using neural networks.

Said Thomas E. Berghage, co-founder and CEO of NeuWorld Financial, San Diego: “We would love to see more people in this game. Several big players are needed to help spread the word with regard to artificial intelligence.”

After reviewing the American Century filing documents, however, Mr. Berghage said Newton may not be completely in the artificial intelligence camp. “It looks like they may be using some human rules, not total artificial intelligence,” he said, because the fund will use a computerized process to screen stocks with a formula based on how its investment managers use technical indicators to select stocks, not “pure” AI.

NeuWorld uses a “pure” artificial intelligence program written by Mr. Berghage to monitor 4,000 domestic stocks, he said. “We take the human element out of the process.”

The system analyzes 116 variables on each stock daily and makes buy and sell recommendations. A typical Wall Street analyst covers 30-40 companies and can only use up to a dozen variables accurately, said NeuWorld portfolio manager Barry Hippensteel. The NeuWorld portfolio, which includes about 110 domestic stocks, gained 30% in the first quarter and is up 15% during the first half of this year. In the year ended June 30, the portfolio was up 36.97%, he said.

Good Wishes

“I hope they do well. Clearly they have a large marketing budget, and it will be good for artificial intelligence and those in the industry who use it,” said Barry Schneirov, managing director at Advanced Investment Technology Inc., Clearwater, Fla., a quantitative equity manager and member of State Street Global Alliance LLC.

AIT uses neural networks as a tool in its stocks analysis and selection process. Mr. Schneirov said AIT uses artificial intelligence as one of several analytical tools, which “gives us a perspective that is different than 99% of our peers.”

Lou Mendelsohn, president of Market Technologies, a Wesley Chapel, Fla.- based firm that designs and tests neural trading systems for the financial industry, said artificial intelligence is no longer just a “buzz word”. Recent technological developments have enhanced the reliability of neural systems, he said, and the trading systems developed by Market Technologies are capable of predicting short-term price movements with 70% to 80% accuracy, he said.

Nevertheless, said Mr. Mendelsohn, AI is “not the holy grail and not an end itself.”

“Markets aren’t 100% predictable; there is still a randomness and unpredictability involved,” he said. “It’s a question of, at the margin, do you have an edge? If artificial intelligence can be used to provide even a slight edge, it offers value added. Anything over 50% is an edge.”

Mr. Mendelsohn said artificial intelligence had made great strides since initial efforts to use the technology in the financial services industry in the 1970’s and 1980’s. Computing power has increased by several orders magnitude in the last 10 years alone, he said, and it “can support the types of number crunching it takes to make use of neural networks.”