基于Seq2Seq文本摘要和情感挖掘的股票波动趋势预测

管理评论 ›› 2021, Vol. 33 ›› Issue (5): 257-269.

• 物理-事理-人理系统方法论（WSR） • 上一篇下一篇

基于Seq2Seq文本摘要和情感挖掘的股票波动趋势预测

齐甜方, 蒋洪迅

中国人民大学信息学院, 北京 100872

收稿日期:2018-10-22 出版日期:2021-05-28 发布日期:2021-06-03
通讯作者: 蒋洪迅(通讯作者),中国人民大学信息学院副教授,博士
作者简介:齐甜方,中国人民大学信息学院硕士研究生。
基金资助:
国家自然科学基金项目（71571183）。

Exploring Stock Price Trend Using Seq2Seq Based Automatic Text Summarization and Sentiment Mining

Qi Tianfang, Jiang Hongxun

School of Information, Renmin University of China, Beijing 100872

Received:2018-10-22 Online:2021-05-28 Published:2021-06-03

摘要/Abstract

摘要： 文本情感挖掘技术被广泛应用于金融市场中股票价格预测和波动趋势分析。以往的研究多局限于全文情感挖掘，即把全文各部分视作同等重要，然而对于互联网极具自由精神的新闻文本来说这是一个不切实际的强假设，有些新闻局部内容甚至与主题无关。本文提出一种基于自动文本摘要和情感挖掘技术的股票波动趋势预测研究框架，将Seq2Seq文本摘要技术应用于情感挖掘中，来提升文本情感表达的准确性，之后将新闻摘要情感值作为股票预测模型的输入特征。为了验证模型有效性，本文进行了新闻摘要情感值和全文情感值预测的交叉对比实验，结果表明二者情感波动趋势大多数情况下趋于一致，但是文本摘要的波动区间更小而且分布也更为稳定；在情感值分布上，摘要挖掘对负向情感更加敏感；对24只沪市A股历时两年半的涨跌趋势预测方面，摘要情感挖掘模型的预测效果要明显优于全文情感挖掘方法。

关键词: 股票涨跌幅, 自动文本摘要, Seq2Seq, 情感分析

Abstract: Text sentiment mining is widely used in empirical studies for stock prices forecasting or trend analysis. Existing researches tend to focus merely on mining the full texts of financial news with an underlying assumption that all contents are related to its topic and all sections count equally. Clearly this is unduly restrictive due to the presence of irregular text format and structural confusion of internet news. This paper proposes and studies stock trend prediction models with text summarization and sentiment mining. We use Seq2Seq method to summarize news texts automatically, then mine their sentiment to improve the performance of models, and finally incorporate these sentiment values as additional features into the machine learning of stock prediction. We conduct experiments to compare two strategies, one with sentiment values of full texts and the other with sentiment value of text summary. The results show that a) their curves of emotional fluctuation are the same in most cases, but the one with text summary has smaller range and is more stable; b) the one with text summary is more sensitive to negative emotional words; c) the one with text summary, in the stock market trend prediction, obviously dominates the other.

Key words: stock price trend, automatic text summarization, Seq2Seq, sentiment analysis

齐甜方, 蒋洪迅. 基于Seq2Seq文本摘要和情感挖掘的股票波动趋势预测[J]. 管理评论, 2021, 33(5): 257-269.

Qi Tianfang, Jiang Hongxun. Exploring Stock Price Trend Using Seq2Seq Based Automatic Text Summarization and Sentiment Mining[J]. Management Review, 2021, 33(5): 257-269.

参考文献

[1] 夏楸,杨一帆,郑建明. 媒体报道、媒体公信力与债务成本[J]. 管理评论, 2018,30(4):180-193
[2] 王春峰,李思成,房振明. 投资者认知度、信息不对称与股价延迟[J]. 管理评论, 2018,30(11):3-13
[3] 金秀,姜尚伟,苑莹. 基于股吧信息的投资者情绪与极端收益的可预测性研究[J]. 管理评论, 2018,30(7):16-25
[4] 尹海员,华亦朴. 我国股票市场流动性的非线性动力学特征研究:基于分形理论的检验[J]. 管理评论, 2017,29(8):43-52
[5] 孙瑞奇. 基于LSTM神经网络的美股股指价格趋势预测模型的研究[D]. 首都经济贸易大学硕士学位论文, 2015
[6] Patel J., Shah S., Thakkar P., et al. Predicting Stock and Stock Price Index Movement Using Trend Deterministic Data Preparation and Machine Learning Techniques[J]. Expert Systems with Applications An International Journal, 2015,42(1):259-268
[7] Hassan M. R., Nath B., Kirley M. A Fusion Model of HMM, ANN and GA for Stock Market Forecasting[J]. Expert Systems with Applications, 2007,33(1):171-180
[8] Ticknor J. L. A Bayesian Regularized Artificial Neural Network for Stock Market Forecasting[J]. Expert Systems with Applications, 2013,40(14):5501-5506
[9] Kim K. J. Financial Time Series Forecasting Using Support Vector Machines[J]. Neurocomputing, 2003,55(1):307-319
[10] Liu F., Wang J. Fluctuation Prediction of Stock Market Index by Legendre Neural Network with Random Time Strength Function[J]. Neurocomputing, 2012,83(6):12-21
[11] Kim H. Y., Won C. H. Forecasting the Volatility of Stock Price Index:A Hybrid Model Integrating LSTM with Multiple GARCH-Type Models[J]. Expert Systems with Applications, 2018,103:25-37
[12] 史建楠,邹俊忠,张见,等. 基于DMD-LSTM模型的股票价格时间序列预测研究[J/OL]. 计算机应用研究, https://doi.org/10.19734/j.issn.1001-3695.2018.08.0657, 2019-03-02
[13] 龙奥明,毕秀春,张曙光. 基于LSTM神经网络的黑色金属期货套利策略模型[J]. 中国科学技术大学学报, 2018,48(2):125-132
[14] Klibanoff P., Lamont O., Wizman T. A. Investor Reaction to Salient News in Closed-End Country Funds[J]. Journal of Finance, 2010,53(2):673-699
[15] Li X., Xie H., Chen L., et al. News Impact on Stock Price Return via Sentiment Analysis[J]. Knowledge-Based Systems, 2014,69:14-23
[16] Hagenau M., Liebmann M., Neumann D. Automated News Reading:Stock Price Prediction Based on Financial News Using Context-Capturing Features[J]. Decision Support Systems, 2013,55(3):685-697
[17] Li Q., Wang T. J., Li P., et al. The Effect of News and Public Mood on Stock Movements[J]. Information Sciences, 2014,278(10):826-840
[18] Kahneman D. Attention and Effort[M]. Englewood Cliffs:Prentice-Hall, 1973
[19] Pashler H., Johnston J. C., Ruthruff E. Attention and Performance[J]. Annual Review of Psychology, 2001,52(1):629-651
[20] Campello M., Graham J. R. Do Stock Prices Influence Corporate Decisions? Evidence from the Technology Bubble[J]. Journal of Financial Economics, 2013,107(1):89-110
[21] Mittal A., Goel A. Stock Prediction Using Twitter Sentiment Analysis[DB/OL]. Standford University, CS229(http://cs229.stanford.edu/proj2011/GoelMittalStockMarketPredictionUsingTwitterSentimentAnalysis.pdf, 2011
[22] Fung G. P. C., Yu J. X., Lam W. News Sensitive Stock Trend Prediction[A]. Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining[C]. Springer-Verlag, 2002
[23] Farra N., Challita E., Assi R. A., et al. Sentence-Level and Document-Level Sentiment Mining for Arabic Texts[A]. IEEE International Conference on Data Mining Workshops[C]. IEEE, 2011
[24] 吴江,唐常杰,李太勇,等. 基于语义规则的Web金融文本情感分析[J]. 计算机应用, 2014,34(2):481-485
[25] 李倩,许伟,蒋洪迅. 考虑用户口碑的旅游计划个性化推荐方法研究[J]. 管理评论, 2016,28(6):113-118
[26] 陈卓群,王平. 面向中文微博摘录式摘要方法研究[J]. 情报科学, 2015,(3):130-134
[27] Cao Z., Wei F., Li S., et al. Learning Summary Prior Representation for Extractive Summarization[C]. Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2:Short Papers), 2015
[28] Chopra S., Auli M., Rush A. M. Abstractive Sentence Summarization with Attentive Recurrent Neural Networks[C]. Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies, 2016
[29] 庞超,尹传环. 基于分类的中文文本摘要方法[J]. 计算机科学, 2018,45(1):144-147
[30] Abadi M., Agarwal A., Barham P. TensorFlow:Large-Scale Machine Learning on Heterogeneous Distributed Systems[DB/OL]. https://ui.adsabs.harvard.edu/abs/2016arXiv160304467A, 2016
[31] Bahdanau D., Cho K., Bengio Y. Neural Machine Translation by Jointly Learning to Align and Translate[DB/OL]. https://ui.adsabs.harvard.edu/abs/2014arXiv1409.0473B, 2014
[32] Hochreiter S., Schmidhuber J. Long Short-Term Memory[J]. Neural Computation, 1997,9(8):1735-1780
[33] Hub C., Lcsts Z. A Large Scale Chinese Short Text Summarization Dataset[C]. Procee-dings of the 2015 Conference on Empirical Methods in Natural Language Processing. Lisbon:ACL, 2009
[34] 蔺璜,郭姝慧. 程度副词的特点范围与分类[J]. 山西大学学报(哲学社会科学版), 2003,26(2):71-74
[35] 文贞惠. 现代汉语否定范畴研究[D]. 复旦大学博士学位论文, 2003

基于Seq2Seq文本摘要和情感挖掘的股票波动趋势预测

Exploring Stock Price Trend Using Seq2Seq Based Automatic Text Summarization and Sentiment Mining

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 3

编辑推荐

Metrics

本文评价

[1]	张振刚, 罗泰晔. 基于在线评论数据挖掘和Kano模型的产品需求分析[J]. 管理评论, 2022, 34(11): 109-117.
[2]	赵宇晴, 阮平南, 刘晓燕, 单晓红. 基于在线评论的用户满意度评价研究[J]. , 2020, 32(3): 179-189.
[3]	孟园, 王洪伟, 王伟. 网络口碑对产品销量的影响:基于细粒度的情感分析方法[J]. , 2017, 29(1): 144-154.