基于粒子群优化LSTM的股票预测模型

被引:267
作者
宋刚 [1 ,2 ]
张云峰 [1 ,2 ]
包芳勋 [3 ]
秦超 [1 ,2 ]
机构
[1] 山东财经大学计算机科学与技术学院
[2] 山东财经大学山东省数字媒体技术重点实验室
[3] 山东大学数学学院
关键词
粒子群优化(PSO); LSTM神经网络; 自适应; 股票价格预测; 预测精度;
D O I
暂无
中图分类号
TP18 [人工智能理论]; F830.91 [证券市场];
学科分类号
020219 [财政学(含:税收学)]; 140502 [人工智能];
摘要
为了提高股票时间序列预测精度,增强预测模型结构参数可解释性,提出一种基于自适应粒子群优化(PSO)的长短期记忆(LSTM)股票价格预测模型(PSO-LSTM),该模型在LSTM模型的基础上进行改进和优化,因此擅长处理具有长期依赖关系的、复杂的非线性问题。通过自适应学习策略的PSO算法对LSTM模型的关键参数进行寻优,使股票数据特征与网络拓扑结构相匹配,提高股票价格预测精度。实验分别以沪市、深市、港股股票数据构建了PSO-LSTM模型,并对该模型的预测结果与其他预测模型进行比较分析。结果表明,基于自适应PSO的LSTM股票价格预测模型不但提高了预测准确度,而且具有普遍适用性。
引用
收藏
页码:2533 / 2542
页数:10
相关论文
共 11 条
[1]
一种更简化而高效的粒子群优化算法 [J].
胡旺 ;
李志蜀 .
软件学报, 2007, (04) :861-868
[2]
Deep learning with long short-term memory networks for financial market predictions[J] Thomas Fischer;Christopher Krauss European Journal of Operational Research 2018,
[3]
Introduction to the special issue on deep learning approaches for machine translation[J] Marta R. Costa-jussà;Alexandre Allauzen;Loïc Barrault;Kyunghun Cho;Holger Schwenk Computer Speech & Language 2017,
[4]
Evaluating deep learning architectures for Speech Emotion Recognition[J] Haytham M. Fayek;Margaret Lech;Lawrence Cavedon Neural Networks 2017,
[5]
Diagnosing deep learning models for high accuracy age estimation from a single image[J] Junliang Xing;Kai Li;Weiming Hu;Chunfeng Yuan;Haibin Ling Pattern Recognition 2017,
[6]
Maxout neurons for deep convolutional and LSTM neural networks in speech recognition[J] Meng Cai;Jia Liu Speech Communication 2016,
[7]
Recurrent neural network and a hybrid model for prediction of stock returns[J] Akhter Mohiuddin Rather;Arun Agarwal;V.N. Sastry Expert Systems With Applications 2015,
[8]
OBST-based segmentation approach to financial time series[J] Yain-Whar Si;Jiangling Yin Engineering Applications of Artificial Intelligence 2013,
[9]
Short-term electric load and temperature forecasting using wavelet echo state networks with neural reconstruction[J] Ali Deihimi;Omid Orang;Hemen Showkati Energy 2013,
[10]
Long Short-Term Memory[J] Sepp Hochreiter;Jürgen Schmidhuber Neural Computation 1997,