A Stock Market Prediction System Based on High-Level Fuzzy Petri Nets

Abstract

As information technology advances dramatically and the stock market in Taiwan turns out to be active, the management of financial investment has already become very important. To facilitate the rapid development of the stock market, the application of information technology to financial investment becomes a hot issue. In this paper, a support vector regression (SVR) machine for stock prices in Taiwan is adopted to simulate the approximate trading trend by using the daily data sets. Then the learned data model is used to analyze technical indices, to draw a trend diagram, and to make a prediction. Finally, the business behavior of financial investment systems has been modeled by using a high-level fuzzy Petri net (HLFPN) for the purpose of making a decision on appropriate investment. Based on the HLFPN model, the proposed approach provides each investor with relevant information to understand the investment trend. As a result, a practical market stock prediction system is presented to enhance the investment benefits and help investors achieve the desired investment goal.

A modular Takagi-Sugeno-Kang (TSK) system based on a modified hybrid soft clustering for stock selection

“…Traders use this approach for short-term investment horizons. They deal with the stock timing and not the company's financial health [3][4][5][6][7]. Since the fundamentals have a stronger relationship to the price movement in the longer horizons, the stock selection stage should be designed based on the fundamental analysis.…”