Devin Liang

論文名稱	An Advanced Stock Pattern Recognition Model Based on A Break-Point Search Algorithm to Forecast DJIA Index
研討會開始日期	2019-07-19
研討會結束日期	2019-07-22
所有作者	Tai-Liang Chen, Ting-Wen Liang
作者順序	第二作者
通訊作者	否
研討會名稱	The International Conference on Innovative Computing and Management Science (ICMS 2019)
是否具有對外公開徵稿及審稿制度	是
研討會舉行之國家	NATJPN-日本
研討會舉行之城市	OSAKA
發表年份	2019
所屬計劃案	無
可公開文檔
可公開文檔
可公開文檔

[摘要] :
NONE

[英文摘要] :
Stock market forecasting is a popular research area for academic researchers. Many researchers proposed
forecasting models with advanced algorithms (i.e., neural networks, fuzzy time-series, and hidden Markov models) to generate
a forecast for a stock price or index and evaluated them with forecasting error (e.g., root-mean-square error, RMSE). In addition
to these models, other researchers issued a different way to forecast stock markets by stock pattern recognizing and proposed
many stock pattern recognition models. In this paper, I propose an advanced stock pattern recognition model, which provides
a break-point (the local minimum of a 20-day stock pattern) search algorithm and fuzzy logical relationships (FLRs) of technical
indicators to discover the signal for the break-point in the stock market. To verify the proposed model, I employ the 12-year
period of the Dow Jones Industrial Average (DJIA), from 2007 to 2018, as experimental datasets and a sliding window
simulation process to implement the experiments with different ratios of training to testing periods. Besides, I take four trading
strategies (5-day, 10-day, 15-day and 20-day holding periods) to evaluate the proposed model by stock index returns. From the
experimental results, the proposed model has given a high average of profitable transaction rate (70%) when the ratio of training
to testing periods is set as 6:1. The averages of profitable stock index return (from 2013 to 2018) for four trading strategies are
4449 (5-day), 7696(10-day), is 11796(15-day), and 11752(20-day). In the initial verification, the proposed model can find the
break-points in the DJIA with a high probability and it seems that there are clear signals for them.

[參考文獻] :
Arévalo, R., García, J., Guijarro, F., & Peri, A. (2017). A dynamic trading rule based on filtered flag pattern recognition for
stock market price forecasting. Expert Systems with Applications, 81, 177-192.
Chen, T. L., & Chen, F. Y. (2016). An intelligent pattern recognition model for supporting investment decisions in
stock market. Information Sciences, 346, 261-274.
Duda, R. O., & Hart, P. E. (1973). Pattern classification and scene analysis. New York: John Wiley & Sons.
Lee, C.-F., Lee, J. C., Lee, A. C. (2013). Statistics for Business and Financial Economics (3rd Eds.). New York: Springer.
Leigh, W., Purvis, R., & Ragusa, J. M.(2002). Forecasting the NYSE composite index with technical analysis, pattern
recognizer, neural networks, and genetic algorithm: a case study in romantic decision support. Decision Support Systems, 32,
361-377.
Pal, S. S., & Kar, S. (2019). Time series forecasting for stock market prediction through data discretization by fuzzistics and
rule generation by rough set theory. Mathematics and Computers in Simulation, 162, 18-30.
Wang, J. L., & Chan, S. H. (2007). Stock market trading rule discovery using pattern recognition and technical analysis. Expert
Systems with Applications, 33 (2), 304-315.
Zhang, M., Jiang, X., Fang, Z., Zeng, Y., & Xu, K. (2019). High-order Hidden Markov Model for trend prediction in
financial time series. Physica A: Statistical Mechanics and its Applications, 517, 1-12.
Zhou, F., Zhou, H., Yang, Z., & Yang, L. (2019). EMD2FNN: A strategy combining empirical mode decomposition and
factorization machine based neural network for stock market trend prediction. Expert Systems with Applications, 115,136-151.