DZone
Big Data Zone
Thanks for visiting DZone today,
Edit Profile
  • Manage Email Subscriptions
  • How to Post to DZone
  • Article Submission Guidelines
Sign Out View Profile
  • Post an Article
  • Manage My Drafts
Over 2 million developers have joined DZone.
Log In / Join
  • Refcardz
  • Trend Reports
  • Webinars
  • Zones
  • |
    • Agile
    • AI
    • Big Data
    • Cloud
    • Database
    • DevOps
    • Integration
    • IoT
    • Java
    • Microservices
    • Open Source
    • Performance
    • Security
    • Web Dev
DZone > Big Data Zone > Deep Learning and the Stock Market

Deep Learning and the Stock Market

Learn how Deep Learning is affecting the stock market and what Artificial Intelligence has to do with capital market data.

Adi Gaskell user avatar by
Adi Gaskell
·
Apr. 20, 17 · Big Data Zone · Analysis
Like (1)
Save
Tweet
7.73K Views

Join the DZone community and get the full member experience.

Join For Free

The financial industry has been one of the more enthusiastic exponents of automation in recent years — especially on the trading floor, where algorithmic trading is now commonplace. Indeed, Commerzbank has been outspoken about its desire to automate 80% of its processes in the coming years.

A sign of the progress being made comes via a recent paper published by the School of Business and Economics at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). It shows the potential for algorithms to make extremely profitable investment decisions.

When the algorithm was applied to members of the S&P 500 between 1992 and 2015, the picks generated double-digit annual returns, with a particular strength during times of economic turmoil.

AI and Capital Market Data

“Equity markets exhibit complex, often non-linear dependencies,” the team say. “However, when it comes to selecting stocks, established methods are mainly modeling simple relationships. For example, the momentum effect only focuses on a stock’s return over the past months and assumes a continuation of that performance in the months to come. We saw the potential for improvements.”

The researchers used a combination of Deep Learning, gradient boosting, and random forests to generate daily predictions for every stock in the S&P 500 during the time period.

Each method was trained using 180 million different data points, with the algorithms developing knowledge of a complex function that described the relationship between the price-based features of the stock, and its future performance, with impressive results.

“Since the year 2000, we observed statistically and economically significant returns of more than 30% per annum. In the nineties, results were even higher, reflecting a time when our machine learning approaches had not yet been invented,” the authors say.

Overcoming the Efficient Market

The authors believe their work poses a considerable challenge to the efficient-market hypothesis with returns from the algorithm particularly strong during turbulent periods such as the dot-com crash or the global credit crunch in 2008. The authors believe this is because such periods tend to see emotion driving investment decisions.

While the results are undoubtedly promising, the researchers urge a degree of caution, and not to go overboard in any belief that such an approach represents a holy grail.

“In the latter years of the study profitability fell and even dipped into the negative at times. We assume that this decline was driven by the rising influence of artificial intelligence in modern trading – facilitated by increasing computing power and the popularization of Machine Learning. We assume that the downturn may have been caused by the increasing influence of Artificial Intelligence in modern trading,” they say.

They’re looking to refine the process however by feeding the algorithms with much bigger data sets and deeper network architectures to help identify temporal dependencies. In early tests of this modified approach, the forecasting ability and quality rose significantly, even against modern algorithmic peers. It will be a project to watch with interest.

Deep learning AI

Published at DZone with permission of Adi Gaskell, DZone MVB. See the original article here.

Opinions expressed by DZone contributors are their own.

Popular on DZone

  • Introduction to Word Vectors
  • Data Mesh — Graduating Your Data to Next Level
  • How to Get GDPR and Customer Communications Right
  • Python 101: Equality vs. Identity

Comments

Big Data Partner Resources

X

ABOUT US

  • About DZone
  • Send feedback
  • Careers
  • Sitemap

ADVERTISE

  • Advertise with DZone

CONTRIBUTE ON DZONE

  • Article Submission Guidelines
  • MVB Program
  • Become a Contributor
  • Visit the Writers' Zone

LEGAL

  • Terms of Service
  • Privacy Policy

CONTACT US

  • 600 Park Offices Drive
  • Suite 300
  • Durham, NC 27709
  • support@dzone.com
  • +1 (919) 678-0300

Let's be friends:

DZone.com is powered by 

AnswerHub logo