基于LSTM的股票价格的多分类预测
Multi-Category Prediction of Stock Price Based on LSTM
DOI: 10.12677/CSA.2019.910211, PDF,  被引量   
作者: 陈奉贤*：同济大学，电子与信息工程学院，上海；赵建群：广东科贸职业学院，经济管理学院，广东 广州
关键词: 股票预测；LSTM；沪深300；多分类；Stock Forecasting； LSTM； Shanghai and Shenzhen 300； Multiple Classification
摘要: 为对股票价格的涨跌幅度进行预测，提出一种基于长短期记忆网络(LSTM)的方法。根据股票涨跌幅问题，通过对涨跌幅度做多值量化分类，将其转化成一个多分类问题。将股票的基本交易信息作为特征输入，利用神经网络对其训练，最后对股票的涨跌幅度做分类预测。数据集分沪深300成分股整体类、银行类和证券类三种股票集，实验结果表明该模型在涨跌幅多分类情况下，有比较好的预测效果；同时，在对某一类股票进行预测时，用该类股票的历史交易信息训练的模型要比以整体股票交易信息训练的模型效果好。
Abstract: In order to predict the rise and fall of stock prices, a method based on long-term and short-term memory network (LSTM) is proposed. According to the stock price increase and decrease, through the quantitative classification of the ups and downs, it is transformed into a multi-classification problem. The basic transaction information of the stock is used as the feature input, and it is trained by the neural network. Finally, the stock’s ups and downs are classified and predicted. The data set is divided into three parts: the whole category set of the Shanghai and Shenzhen 300 constituent stocks, the bank set and the securities. The experimental results show that the model has a better forecasting effect in the case of multi-classification of the ups and downs; at the same time, in a certain kind of stocks when making predictions, the model trained with the historical trading infor-mation of such stocks is better than the model trained with the overall stock trading information.
文章引用：陈奉贤, 赵建群. 基于LSTM的股票价格的多分类预测[J]. 计算机科学与应用, 2019, 9(10): 1882-1891. https://doi.org/10.12677/CSA.2019.910211

参考文献

[1] 刘长虎, 陶建格, 崔衍秋. 股票价格指数的投资功能[J]. 市场论坛, 2004(6): 71-72.
[2] 黄丽明, 陈维政, 闫宏飞. 基于循环神经网络和深度学习的股票预测方法[J]. 广西师范大学学报(自然科学版), 2019, 137(1): 13-22.
[3] 沈金榕. 基于决策树的逐步回归算法及在股票预测上的应用[D]: [硕士学位论文]. 广州: 广东工业大学, 2017.
[4] Huo, J., Zheng, Y. and Chen, X. (2009) Implementation of Transaction Trend Prediction Model Based on Regression Analysis. Journal of Baosha Teachers’ College, 117, 19-23.
[5] 黄秋萍, 周霞, 甘宇健, 韦宇. SVM与神经网络模型在股票预测中的应用研究[J]. 微型机与应用, 2015, 34(5): 88-90.
[6] 刘磊. 基于深度学习的股票价格趋势预测方法研究[D]: [硕士学位论文]. 昆明: 云南财经大学, 2017.
[7] 欧阳金亮, 陆黎明. 综合改进BP神经网络算法在股价预测中的应用[J]. 计算机与数字工程, 2011, 39(2): 57-59.
[8] 毛景慧. 基于LSTM深度神经网络的股市时间序列预测精度的影响因素研究[D]: [硕士学位论文]. 广州: 暨南大学, 2017.
[9] 阎平凡, 张长水. 人工神经网络与模拟进化计算[M]. 北京: 清华大学出版社, 1900.
[10] 闻新, 张兴旺, 朱亚萍, 李新. 智能故障诊断技术: MATLAB应用[M]. 北京: 北京航空航天大学出版社, 2015.
[11] Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., Liu, T., Wang, X., Wang, L., Wang, G. and Cai, J. (2015) Recent Advances in Convolu-tional Neural Networks.
[12] Goodfellow, I., Bengio, Y. and Courville, A. (2016) Deep Learning (Vol. 1). MIT Press, Cambridge, 326-366.
[13] Graves, A. and Schmidhuber, J. (2005) Framewise Phoneme Classification with Bidirectional LSTM and Other Neural Network Architectures. Neural Networks, 18, 602-610. [Google Scholar] [CrossRef] [PubMed]
[14] KIMY (2014) Convolutional Neural Networks for Sentence Classification. In: Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, ACL, Stroudsburg, 1746-1751. [Google Scholar] [CrossRef] 
[15] Kingma, D.P. and Bal, J. (2014) Adam: A Method for Stochastic Opti-mization. Computer Science.
https://arXiv.org/abs/1412.6980