1. Intelligent Database Systems Lab N.Y.U.S.T. I. M. A quantitative stock prediction system based on financial news Presenter : Chun-Jung Shih Authors :Robert P. Schumaker, Hsinchun Chen IPM 2009 國立雲林科技大學 National Yunlin University of Science and Technology 1

2. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Outline Motivation Objective Methodology Experiments Conclusion Comments 2

3. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Motivation Predicting changes in the stock market has always had a certain appeal to researchers. Acquiring relevant textual data is an important facet of stock market prediction. 3

4. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Objective To create the Arizona Financial Text System (AZFinText) Seeks to contribute to the AZFinText system by comparing AZFinText’s predictions against existing quantitative funds and human stock pricing experts. 4 2317 鴻海

5. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology 5

6. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology Textual analysis  To identify the Proper Nouns  Use Arizona Text Extractor (AzTeK) system Stock Quotations  Gathers stock price data in 1 min increments Model Building  Provide superior performance to all combinations tested Trading Experts  Gathers the daily buy/sell recommendations from a variety of trading experts 6

7. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology Metrics  Evaluates system output 7  Closeness  Directional Accuracy  Simulated Trading

8. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Experiments 8

9. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Experiments 9

10. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Experiments 10

11. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Experiments 11

12. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Conclusion 12 Sector had the best Directional Accuracy at 71.18% and Simulated Trading of 8.50% return on investment. Sector also had the second-lowest Closeness score, 0.1954, as compared to Universal, 0.0443. AZFinText had a Directional Accuracy of 71.18%, which was second-best to DayTraders.com’s 81.82%.

13. Intelligent Database Systems Lab N.Y.U.S.T. I. M. Comments 13 Advantage  Predicting changes in the stock market Drawback  DayTraders.com’s Directional Accuracy batter than AZFinText Application  Information Retrieval