Stock Market Trend Prediction Based on Dynamic Influence Diagrams

doi:10.3778/j.issn.1002-8331.2101-0050

Abstract

Abstract: The influence of historical data on the future state is hidden, which makes the stock market trend prediction based on data a difficult problem. In order to effectively discover the influence of historical data on the future state of the stock market, dynamic influence diagrams are used to model the structural relationship between trading volume and K-line patterns, and a joint tree inference prediction algorithm based on the structural relationship between volume and price structure is proposed（VP-JT）. This paper extracts the characteristics of phase volume and phase K-line morphological features of the stock, and gives the principle of the effect of phase volume on the stock market price. The degree of cooperation is used to quantify the consistency of the relationship between the current phase volume and the K-line pattern. Furthermore, the dynamic influence diagram is used to model the action process of volume and price on time in the modeling stage. The future trend of the stock market is predicted by the automatic inference of the joint tree. After the algorithm being implemented and compared on actual data, it is shown by the experimental results that the joint tree inference algorithm based on the structural relationship between volume and price has higher accuracy.

Key words: dynamic influence diagrams, degree of cooperation, utility function, junction tree

摘要： 历史数据对未来状态的影响具有隐蔽性，导致基于数据的股市趋势预测是一个公开难题。为了有效地发现历史数据对股市未来状态的影响力，利用动态影响图建模成交量和Ｋ线形态之间的结构关系，提出一种基于量价结构关系的联合树推理预测算法（VP-JT）。提取股票的阶段成交量特征和阶段Ｋ线形态特征，给出阶段成交量对于股市价格影响的作用原理；利用配合度量化当前阶段成交量与Ｋ线形态之间关系一致性程度；利用动态影响图建模阶段量价在时间上的作用过程；通过联合树的自动推理对股市未来状态进行预测。在实际数据上进行实现和算法比较，实验结果表明量价结构关系的联合树推理算法具有更高的准确率。

关键词: 动态影响图, 配合度, 效用函数, 联合树

YAO Hongliang, XU Liwei, YANG Jing, YU Kui. Stock Market Trend Prediction Based on Dynamic Influence Diagrams[J]. Computer Engineering and Applications, 2022, 58(15): 162-168.

姚宏亮, 徐礼维, 杨静, 俞奎. 基于动态影响图的股市趋势预测[J]. 计算机工程与应用, 2022, 58(15): 162-168.

References

[1] HENRIQUE B M，SOBREIRO V A，KIMURA H.Literature review：machine learning techniques applied to financial market prediction[J].Expert Systems with Applications，2019，124（6）：226-251.
[2] ZUO Y，KITA E.Stock price forecast using Bayesian network[J].Expert Systems with Applications，2012，39（8）：6729-6737.
[3] CHENG S H.A hybrid predicting stock return model based on bayesian network and decision tree[C]//International Conference on Industrial，Engineering and Other Applications of Applied Intelligent Systems.Cham：Springer，2014：218-227.
[4] LI Z，YANG J，TAN S.Systematically discovering dependence structure of global stock markets using dynamic bayesian network[J].Journal of Computational Information Systems，2014，9（18）：7215-7226.
[5] ZHANG M，JIANG X，FANG Z，et al.High-order hidden Markov model for trend prediction in financial time series[J].Physica A：Statal Mechanics and Its Applications，2018，517.
[6] ZHAO Z，RAO R，TU S，et al.Time-weighted LSTM model with redefined labeling for stock trend prediction[C]//2017 IEEE 29th International Conference on Tools with Artificial Intelligence（ICTAI），2017：1210-1217.
[7] WEN M，LI P，ZHANG L，et al.Stock market trend prediction using high-order information of time series[J].IEEE Access，2019，7：28299-28308.
[8] HOOI B，LIU S，SMAILAGIC A，et al.“BeatLex：summarizing and forecasting time series with patterns[C]//The European Conference on Machine Learning & Principles and Practica of Knowledge Discovery in Databases，Skopje，Macedonia，2017.
[9] 孙彦林，陈守东，刘洋.基于股市和汇市成交量信息视角的股价波动预测[J].系统工程理论与实践，2019，39（4）：935-945.
SUN Y L，CHEN S D，LIU Y.Forecast of stock price fluctuation based on the perspective of volume information in stock and foreign exchange market[J].System Engineering-Theory & Practice，2019，39（4）：935-945.
[10] MITCHELL D，BIAKOWSKI J，TOMPAIDIS S.Volume-weighted average price tracking：a theoretical and empirical study[J].IISE Transactions，2019，51（8）：1-26.
[11] AMINIKHANGHAHI S，COOK D J.A survey of methods for time series change point detection[J].Knowledge and Information Systems，2017，51（2）：339-367.
[12] MALAGRINO L S，ROMAN N T，MONTEIRO A M.Forecasting stock market index daily direction：a Bayesian network approach[J].Expert Systems with Applications，2018，105（9）：11-22.
[13] SANDOVAL J，HERNáNDEZ G.Learning of natural trading strategies on foreign exchange high-frequency market data using dynamic Bayesian networks[J].Machine Learning and Data Mining in Pattern Recognition，2014：408-421.
[14] FENG X S，GU J，GUAN X F，et al.Optimal allocation of hybrid energy storage for microgrids based on multi-attribute utility theory[J].Journal of Modern Power Systems & Clean Energy，2018：107-117.
[15] 姚宏亮，王浩，张佑生，等.多Agent动态影响图及其一种近似推理算法研究[J].计算机学报，2008，31（2）：236-244.
YAO H L，WANG H，ZHANG Y S，et al.Research on multi-agent dynamic influence diagrams and its approximate inference algorithm[J].Chinese Journal of Computers，2008，31（2）：236-244.