You are here:  Home  >  Data Education  >  Big Data News, Articles, & Education  >  Current Article

Professional Investors: Get Your Quant In A Box

By   /  December 3, 2012  /  No Comments

The talk of falling off the fiscal cliff that’s drowning out the holiday music could take its toll on what historically is a strong month for the stock market, according to Reuters, How that scenario will play out, not to mention a ton of other factors, is just the kind of thing to keep hedge fund managers, wealth advisors and advanced individual investors on their toes as they calculate investment strategies.

A new cloud-based artificial intelligence solution from Lucena, the first features of which are going live today, focuses on helping these users scientifically validate their investment plans, the idea being to find new market opportunities and reduce risk. The early stage company is headed up by serial entrepreneur and CEO Erez Katz, whose partner in the venture is CTO Tucker Balch, a professor of Interactive Computing at the Georgia Institute of Technology whose work focuses on machine learning and robotics.

QuantDesk is the result of five years of research Balch has done at the institution. It is, as Katz describes it, a “quant in a box” that can give sophisticated investment professionals in small or mid-size firms, who lack the resources of the large investment houses to hire quantitative analysts to derive complex and sophisticated trading algorithms, access to a scientific approach to “validate or pivot the decision process,”

“Big hedge funds have the ability to hire the expensive quants who figure out trends and patterns in the market, looking at the historical behavior of stocks and equities and trying to correlate events or behaviors to future impact,” Katz says. “We do the same, but at lower cost,” with its web-enabled solution that’s built on statistical machine learning technology.

“Machine learning is a complex word but it means that the model knows to adjust itself automatically as conditions in the field change. It’s all driven by pure data, not emotions or non-scientific measures.” The approach should provide investor professionals with a way to augment their decision processes and create a more predictable investment strategy, he says.

There are five pieces to the solution. The Price Forecaster, Portfolio Optimizer, and Hedge Finder are available now. Pattern recognition technology helps identify historical patterns in the Price Forecaster, for example, makes it possible to discover how often a stock was in a similar state to its current one (not a task for mere humans); when QuantDesk finds matches, it can evaluate how the stock behaved towards the future – 5 to 20 days out – from the time it had the match. “Say we found 100 matches in the past [for Apple] and 90 percent of the time Apple went up when you had that closely matched state. Then [QuantDesk] would have a very strong opinion that it goes up again in the next 5, 10 and 20 days,” says Katz.

It computes more than 120 technical indicators from price volume and fundamental data for use by the models, regularly determining based on historical analysis which are the dozen or so most relevant to apply but also letting users choose their own indicators based on their requirements. “We build the model with just 10 or 15 of them or else it becomes counterproductive,” he says. “Things change. Today people are afraid of the fiscal cliff. Once that is over other things will come into the mix of decisions and our models identify that psychological change in the community and address what matters.”

Its Portfolio Optimizer uses the approach developed by Nobel Prize winner Harry Markowitz in order to help managers allocate funds in the most appropriate ways to achieve their clients’ goals. But what makes it unique, Katz says, is that it funnels the future predicted price of a stock determined by the proprietary machine learning algorithm from the Price Forecaster module. With the help of its machine learning algorithm. Hedge Finder aims to reduce volatility, finding candidate equities for the hedge by evaluating all actively traded securities, assessing up to millions of combinations of equities to find the right basket, the company says. Hedge Finder takes a portfolio with certain holdings and lets the user exclude certain equities from the hedge that would otherwise be part of the hedge basket.

Also predicated on its machine learning technology and pattern recognition approach are Event Study, which lets users look at events that are meaningful in the market and how they affect certain equities in ways that are predictable. For instance, Katz says, “if there’s a war with Iran tomorrow, and oil prices go up 30 percent, how would your portfolio or group of equities be affected by it?” Airline holdings, for example, historically have gone down because of the impact on operating costs. “Some of these things are fairly obvious and known but there’s a lot of intrinsic value hidden impact in events to help you identify where to invest in if a certain event occurs in the future,” he says.

With its Back Tester, users also can simulate trading strategies – testing, for instance, a theory that small cap stocks, which may be negatively impacted by an overall market panic, are unfairly punished, and come back strong after the panic subsides. Event Study and Back Tester are due in February.

“People investing 50 to 100 million worth of capital can make a real impact in their success ratios,” Katz says. But don’t count out the larger institutions as potential customers. Even those with their own quant teams, he says, are always looking for investment technology to augment or extend their competitive advantage. Lucene has had discussions with some very large and prominent hedge funds that are interested in a custom augumentation of the technology for their individual use, he notes.

It will take time for professional investors to become comfortable with the idea of a machine learning, AI approach, Katz notes, but the day will come. “This is a hot topic – it’s big data, statistical analysis, and behavioral science. It all comes together in one tool.”





About the author

Jennifer Zaino is a New York-based freelance writer specializing in business and technology journalism. She has been an executive editor at leading technology publications, including InformationWeek, where she spearheaded an award-winning news section, and Network Computing, where she helped develop online content strategies including review exclusives and analyst reports. Her freelance credentials include being a regular contributor of original content to The Semantic Web Blog; acting as a contributing writer to RFID Journal; and serving as executive editor at the Smart Architect Smart Enterprise Exchange group. Her work also has appeared in publications and on web sites including EdTech (K-12 and Higher Ed), Ingram Micro Channel Advisor, The CMO Site, and Federal Computer Week.

You might also like...

Data Science Use Cases

Read More →