构建多层循环神经网络结构对金融指标数据的验证分析

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名

邮箱

手机号码

标题

留言内容

验证码

Buiding Mulit-level Recurrent Neural Network Structure to Analyse Financial Indice Data

1.China United Network Telecommunications Corporation, Nanning Guangxi 530022; 2.Guangxi University of Finance and Economics, Nanning Guangxi 530022

摘要: 我们构建了一种多层循环神经网络来对金融股票指标数据进行测试,分析多种指标算法组合用来预测股票走势的可靠性。在实际的股票分析当中,通常会选择不同的技术指标或者指标组合对趋势进行判断,但在验证这些金融技术指标组合可靠性上需要时间和实践去验证,利用循环神经网络去分析过滤指标数据生成预测数据,通过对预测数据和真实数据对比, 我们可以验证出哪个股票技术指标组合对股票投资决策是可靠的。
- 金融技术指标 /
- 循环神经网络 /
- 损失函数 /
- 神经网络设计 /
- 股票趋势预测
Abstract: We propose a kind of recurrent neural network structure for analysis stock indicators and its composed indicator sets which are used for predicting stock trend. Traditional method to analysis stock trend is using kinds of indicators algorithm, but estimation of effective indicators need long period testing time to evaluate. We compose a multiple rnn cell layers neural network to receive stock data, and output prediction result data basing on sets of indicators’ data which are put through into the network as well. With comparing prediction result and real data, it is helpful to choose composition of stock indicators for investment decision strategy and saving time.
- Financial indicator /
- Recurrent neural network /
- Loss function /
- Development of neural network /
- Stock trend prediction

[1]	冯先成,周密,徐川.基于 Android 的百度地图多功能实现[J].武汉工程大学学报,2016(5):490-494.
[2]	张波,赵双明.基于 Android 平台的百度地图开发研究[J].软件导刊,2015(7):96-99.