Research Spotlight
Smart Market and Money
Hsinchun Chen, Editor and AuthorThree articles presenting unique perspectives, advanced computational methods, selected results and examples of stock market modeling and prediction of market movements.
Published November/December 2011 edition, IEEE Intelligent Systems, IEEEXplore, www.computer.org/intelligent
Market, a term frequently used in mass media and academic publications, is an elusive concept. In marketing, researchers and practitioners describe market as a place to exchange products and services (such as a retail market or real estate market). In economics and finance, financial and monetary concepts such as emerging markets, commodity markets, and the stock market are often mentioned.
In all these areas, one of the
most challenging research directions is modeling
and predicting market movements. In recent
years, the availability of diverse and voluminous
market-related mass media and social media
content (or Business Big Data) and the emergence
of sophisticated, scalable text and social
mining techniques present a unique opportunity
for advancing research relating to smart market
and money. This research area, at the intersection
of computational and finance research, aims
at developing intelligent (smart) mechanisms
and algorithms for predicting market and stock
performances.
This publication includes three article on smart market and money from distinguished experts in information systems and business. Each article presents unique perspectives, advanced computational methods, and selected results and examples.
In "AZ SmartStock: Stock Prediction with Targeted Sentiment and Life Support," Hsinchun Chen, Edward Chun-Neng Huang, Hsinmin Lu, and Shu-Hsing Li report on the design and testing of the AZ Smart-Stock system, which incorporates target sentiment and life support in a prototype stock-trading engine. We considered transaction costs and simulated trading performed using data collected for 129 trading days in 2008. The proposed trading model outperformed other benchmark models in the 10-day trading window. This article also suggests several directions
for future research in predicting market movements.
In "A Stakeholder Approach to Stock Prediction Using Finance Social Media," David Zimbra and Hsinchun Chen describe research that utilizes firm-related finance Web forum discussions to predict stock returns and trading of firm stock. Recognizing the diversity among forum participants, they segmented them into distinct stakeholder groups based on their interactions (posting activities) in the forums. By analyzing finegrained stockholder groups, this system reported improved stock-return prediction versus a baseline system and aggregated forum model.
In the final article, "Computational Intelligence for Smart Markets: Individual Behavior and Preferences," Paulo B. Goes argues that in today's Web-enabled marketplaces, the economic environment is much more complex than the preference modeling used by experimental economists. The monitoring opportunities available with the Internet provide ample opportunities to build analytics and computational intelligence to understand in real time the complexities of the preference structure and behaviors of today's heterogeneous market participants. Goes summarizes selected Web-based auction research that illustrates how to acquire computational intelligence on the preferences and behaviors of the participants in these new microeconomies.
For additional information, please contact us.
