深度学习应用于股票预测 | Deep learning for event-driven stock prediction

本文基于论文 Deep learning for event-driven stock prediction
该论文旨在结合短期、中期和长期新闻事件来共同预测股票价格波动。

文章要点
  1. 利用OpenIE(open information extraction)从新闻文本中抓取事件结构表征(执行器、动作、对象、时间),然后利用神经张量网络或成倍地结合事件参数来学习事件的语义结合性。
    用于事件嵌入的神经张量网络
    o 1 o_1 o1:执行器; o 2 o_2 o2:对象; P P P:动作; T T T:时间)

  2. 利用卷积神经网络结合短期、中期和长期事件共同对股票价格进行预测。基于卷积神经网络的预测模型

关于论文的一些思考

1.在一定程度上,深度学习确实可以帮助我们摸清股市规律。在股市中赚钱是需要掌握信息优势的,需要利用信息的不对称性从信息匮乏的一方赚钱。而深度学习的一大优势就是可以大幅减少人工参与的特征工程去拟合训练数据,好的深度学习策略可以突破人脑的限制从更多维的信息中找到规律。
2.深度学习的作用是有限的。股票的问题不在于计算困难,而在于股市的条件是时刻变化的:小到持股人的出售,大到国家政策的变动和世界经济形势的变化都会对股票价格造成影响。所以将深度学习应用于股市只能作为辅助性工具给炒股者以指导性分析。
3.比模型更重要的是给模型的特征。这篇论文中一个优秀的地方是使用了神经张量网络来学习事件的语义组合性,而不是像标准神经网络那样绝对量地学习,这种方式一定程度上弥补了数据缺失和噪声带来的过拟合缺陷。
4.因为在股市中很多新闻是不透明的,这就会导致数据缺失,给预测增加难度。那么可以考虑从其它类型的财经信号中获取信息,比如黄金、外汇等。还可以将公众对该股票的情感走向纳入考虑范围。

  • 2
    点赞
  • 15
    收藏
    觉得还不错? 一键收藏
  • 1
    评论
We offer a systematic analysis of the use of deep learning networks for stock market analysis and prediction. Its ability to extract features from a large set of raw data without relying on prior knowledge of predictors makes deep learning potentially attractive for stock market prediction at high frequencies. Deep learning algorithms vary considerably in the choice of network structure, activation function, and other model parameters, and their performance is known to depend heavily on the method of data representation. Our study attempts to provides a comprehensive and objective assessment of both the advantages and drawbacks of deep learning algorithms for stock market analysis and prediction. Using high-frequency intraday stock returns as input data, we examine the effects of three unsupervised feature extraction methods—principal component analysis, autoencoder, and the restricted Boltzmann machine—on the network’s overall ability to predict future market behavior. Empirical results suggest that deep neural networks can extract additional information from the residuals of the autoregressive model and improve prediction performance; the same cannot be said when the autoregressive model is applied to the residuals of the network. Covariance estimation is also noticeably improved when the predictive network is applied to covariance-based market structure analysis. Our study offers practical insights and potentially useful directions for further investigation into how deep learning networks can be effectively used for stock market analysis and prediction.
Event-driven programming is a programming paradigm that allows the software to respond to events or user actions in a asynchronous manner. Here are the general steps to implement an event-driven program: 1. Identify the events: Determine the events that the program needs to respond to. These could be user actions, system events, or external signals. 2. Define event handlers: Create functions or methods to handle each event. These functions will be executed when the event occurs. 3. Register event handlers: Associate each event handler with the corresponding event. This is usually done using a framework or library. 4. Run the program: Once the event handlers are registered, the program is ready to execute. As events occur, the associated event handlers will be executed. 5. Clean up: After the program finishes executing, clean up any resources used or allocated during the program. Here's an example of an event-driven program in Python using the Tkinter GUI library: ```python import tkinter as tk def button_click(event): print("Button clicked") root = tk.Tk() button = tk.Button(root, text="Click me") button.bind("<Button-1>", button_click) button.pack() root.mainloop() ``` In this program, we identify the event as a button click, define an event handler function `button_click` to handle the event, register the event handler with the button using the `bind` method, and run the program using the `mainloop` method. When the button is clicked, the `button_click` function will be executed and print "Button clicked" to the console.

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值