[1]马超a,侯天诚b,徐瑾辉a,等.泛函深度神经网络及其在金融时间序列预测中的应用[J].徐州工程学院学报(自然科学版),2017,(2):46-53.
 MA Chaoa,HOU Tianchengb,XU Jinhuia,et al.Application of Functional Deep Neural Network in Financial Time Series Prediction[J].Journal of Xuzhou Institute of Technology(Natural Sciences Edition),2017,(2):46-53.
点击复制

泛函深度神经网络及其在金融时间序列预测中的应用()
分享到:

《徐州工程学院学报》(自然科学版)[ISSN:1674-358X/CN:32-1789/N]

卷:
期数:
2017年第2期
页码:
46-53
栏目:
理论研究
出版日期:
2017-04-28

文章信息/Info

Title:
Application of Functional Deep Neural Network in Financial Time Series Prediction
文章编号:
1674-358X(2017)02-0046-08
作者:
马超12a侯天诚2b徐瑾辉2a3张振华2c4蓝斌2b5
1.伦敦大学学院 计算统计学与机器学习中心,伦敦 WC1E 6BT; 2.广东外语外贸大学 a.金融学院应用数学系,b.金融学院金融学系,c.经济贸易学院统计学系,广东 广州 510006; 3.印第安纳大学 统计学系,布卢明顿 IN 47405; 4.考文垂大学 商务、环境和社会学院,考文垂 CV1 5FB; 5.香港城市大学 经济及金融系,香港 999077
Author(s):
MA Chao12aHOU Tiancheng2bXU Jinhui2a3ZHANG Zhenhua2c4LAN Bin2b5
1.Centre for Computational Statistics and Machine Learning, University College London, London WC1E 6BT, UK; 2a.Department of Applied Mathematics, School of Finance,2b.Department of Finance, School of Finance, 2c.Department of Statistics, School of Econom
关键词:
深度学习 泛函网络 降噪自编码 金融预测
Keywords:
deep learning functional network denoising auto-encoder financial prediction
分类号:
F830.49
文献标志码:
A
摘要:
针对神经网络直接预测原始价格存在的泛化误差大、预测价格变动方向的准确率不高等问题,提出一种基于泛函的深度降噪自编码神经网络,并提高神经网络的在时间序列上的泛化能力.将预测目标改为ZigZag/PI指标,且通过着重预测价格序列的趋势和方向,避免来自原始序列的噪音影响,弥补神经网络在方向预测上的固有缺陷.
Abstract:
To address the problem of low fitting error accompanied with poor generalization and low accuracy in predicting the direction of price movement,a new functional deep auto-encoder neural network(FdAE)is constructed with improved generalization ability.The predicting target is also changed to ZigZag/Position Index(PI),which emphasizes on trend and direction to avoid disturbance of noise from original time series,in order to remedy the inherent deficiency of traditional neural network in direction prediction.

参考文献/References:

[1] BILDIRICI M,ERSIN O.Improving forecasts of GARCH family models with theartificial neural networks:an application to the daily returns in Istanbul stockexchange[J].Expert System with Applications,2009,36:7355-7362.
[2] GODARZI A A,AMIRI R M,TALAEI A,et al.Predicting oil price movements:a dynamic artificial neural network approach[J].Energy Policy,2014,68:371.
[3] AGRAWAL,J G,CHOURASIA,V S,MITTRA A K.State-of-the-art in stock prediction techniques[J].International Journal of Advanced Research in Electrical,2013(2):1360-1366.
[4] HINTON G,DENG L,YU D,etal.Deep neural networks for acoustic modeling in speech recognition:the shared views of four research groups[J].IEEE Signal Processing Magazine,2012,29,82-97.
[5] KUREMOTOT,KIMURA S,KOBAYASHI K,OBAYASHI,M.Time series forecasting using a deep belief network with restricted boltzmann machines[J].Neurocomputing,2014(137):47-56.
[6] HIPPERT H S,PEDREIRA C E,SOUZA R C.Neural networks for short-term load forecasting:a reviewand evaluation[J].IEEE Transactions on Power Systems,2001,16(1):,44-55.
[7] RUGGIERO B,MURRAY A.Neural networks:where are they now?[J].Futures,2006,35(5):52-55.
[8] BENíTEZ-PéREZ H,CáRDENAS-FLORES F,ORTEGA-ARJONA,J L,et al.Fault Localization using neural networks and observers for autonomous elements[J].Control and Intelligent Systems,2007,35(4):319-328.
[9] BISOI R,DASH P K.A hybrid evolutionary dynamic neural network for stock market trend analysis and prediction using unscented kalman filter original research[J].Applied Soft Computing,2014,19:41-56.
[10] HUANG F Y.Integration of an improved particle swarm optimization algo-rithm and fuzzy neural network for Shanghai stock market prediction[C].in:IEEE Workshop on Power Electronics and Intelligent Transportation System,2008.
[11] VINCENT P,LAROCHELLE H,LAJOIE I,et al.Stacked denoising autoencoders:learning useful representations in a deep network with a local denoising criterion[J].Journal of Machine Learning Research,2010,11:3371-3408.
[12] LIU C F,YEH C Y,LEE S J.Application of type-2 neuro-fuzzy modeling in stockprice prediction[J].Applied Soft Computing,2012(12):1348-1358.

备注/Memo

备注/Memo:
收稿日期:2017-03-23 基金项目:广东省大学生科技创新培育专项资金(“攀登计划”专项资金)资助项目(126-GK161012); 国家统计局全国统计科研计划项目(2016LZ18,2016537); 广东省自然科学基金项目(2014A030313575,2016A030313688); 广东省软科学项目(2015A070704051); 广东省质量工程项目(125-XCQ16268); 广东外语外贸大学特色创新及团队项目(15T21,DT1605) 作者简介:马 超(1993-),男,硕士研究生
更新日期/Last Update: 1900-01-01