改进的快速跟踪回声状态网络及PM2.5预测

doi:10.3969/j.issn.1000-1158.2020.09.17

摘要
图/表
参考文献(18)
相关文章 (15)

全文: PDF (1983 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对递归最小二乘回声状态网络在噪声环境中预测精度不高的问题,提出了一种改进的快速跟踪回声状态网络。首先在递归最小二乘回声状态网络结构的基础上,将自适应调节的可变遗忘因子加入其代价函数中,用改进的递归最小二乘法对网络输出权值进行训练,得到快速跟踪回声状态网络;然后利用经典Lorenz混沌系统验证快速跟踪回声状态网络的有效性;最后利用灰关联法分析各相关变量与PM2.5的关联度,建立PM2.5浓度值辅助变量集合,将辅助变量集合输入到快速跟踪回声状态网络进行PM2.5浓度值预测。实验表明,与传统回声状态网络、递归最小二乘回声状态网络预测效果相比,快速跟踪回声状态网络的预测方法精度佳,抗噪声能力强。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	刘彬
	李德健
	赵志彪
	武尤

关键词 ：计量学, PM2.5预测, 回声状态网络, 递归最小二乘法, 灰关联分析

Abstract：Aiming at the problem that the recursive least squares echo state network has low prediction accuracy in noisy environments, an improved fast tracking echo state network was proposed. Firstly, the variable forgetting factor which can be adaptively adjusted was added into the cost function of network, and the output weights of the network were trained by the improved recursive least squares method to obtain a fast tracking echo state network. Then, the effectiveness of the proposed network was verified by classical Lorenz chaotic system. Finally, the grey correlation method was used to analyze the correlation between the relevant variables and PM2.5, and the auxiliary variable set of PM2.5 concentration value was established. The auxiliary variable set was input into the fast tracking echo state network to predict the PM2.5 concentration value. Experimental results show that compared with the traditional echo state network and the recursive least squares echo state network, the improved state echo network has better prediction accuracy and stronger anti-noise ability.

Key words： metrology PM2.5 prediction echo state network recursive least square method grey relational analysis

收稿日期: 2018-10-23 发布日期: 2020-08-28

PACS:

TB99

基金资助:国家自然科学基金(51641609);河北省自然科学基金(E2018203398)

通讯作者: 李德健(1994-),男,黑龙江北安人,燕山大学电气工程学院硕士研究生,主要研究方向为回声状态网络预测控制。 Email: 501807366@qq.com E-mail: liubin@ysu.edu.cn

作者简介: 刘彬(1953-),男,黑龙江哈尔滨人,燕山大学教授,主要从事信号估计与识别算法研究、智能传感与无线传感器网络关键技术研究。Email: liubin@ysu.edu.cn

引用本文:

刘彬,李德健,赵志彪,武尤. 改进的快速跟踪回声状态网络及PM2.5预测[J]. 计量学报, 2020, 41(9): 1138-1145.
LIU Bin,LI De-jian,ZHAO Zhi-biao,WU You. Improved Fast Tracking Echo State Network and PM2.5 Prediction. Acta Metrologica Sinica, 2020, 41(9): 1138-1145.

链接本文:

http://jlxb.china-csm.org:81/Jwk_jlxb/CN/10.3969/j.issn.1000-1158.2020.09.17 或 http://jlxb.china-csm.org:81/Jwk_jlxb/CN/Y2020/V41/I9/1138

［1］Wu J L, Yin X L, Xiao H. Seeing permeability from images: fast prediction with convolutional neural networks［J］. Science Bulletin, 2018, 63(18): 1215-1222.
［2］刘浩然,崔静闯,卢泽丹,等. 一种改进的雁群扩展粒子群算法研究［J］. 计量学报,2019,40(3):498-504.
Liu H R,Cui J C,Lu Z D, et al. Study on an Improved Extended PSO Algorithm Based on Geese Flight. Acta Metrologica Sinica, 2019, 40(3): 498-504.
［3］White D R, Fischer J. The Boltzmann constant and the new kelvin［J］. Metrologia, 2015, 52(5): 213-216.
［4］王有贵, 吴双双, 陈红江. 称重传感器蠕变误差的神经网络补偿方法［J］. 计量学报, 2018, 39(4): 510-514.
Wang Y G, Wu S S, Chen H J. Neural network compensation method for creep error of weighing sensor［J］. Acta Metrologica Sinica, 2018, 39(4): 510-514.
［5］陆艺, 张培培, 王学影. 基于PSO-BP神经网络的关节臂式坐标测量机长度误差补偿［J］. 计量学报, 2017, 38(3): 271-275.
Lu Y, Zhang P P, Wang X Y. Length error compensation of articulated arm CMM based on PSO-BP neural network［J］. Acta Metrologica Sinica, 2017, 38(3): 271-275.
［6］Minkyu Choi, Jun Tani. Predictive Coding for Dynamic Visual Processing: Development of Functional Hierarchy in a Multiple Spatiotemporal Scales RNN Model［J］. Neural Computation, 2017, 30(1): 237-270.
［7］张一珂, 张鹏远, 颜永红. 基于对抗训练策略的语言模型数据增强技术［J］. 自动化学报, 2018, 44(5): 891-900.
Zhang Y K,Zhang P Y,Yan Y H. Language model data enhancement technology based on confrontation training strategy［J］.Automation journal,2018,44(5):891-900.
［8］Jaeger H, Haas H. Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication［J］. Science, 2004, 304(5667): 78-80.
［9］徐正阳, 路志英, 刘洪. 基于泄漏积分型回声状态网络的在线学习光伏功率预测［J］. 电力系统及其自动化学报, 2018, 30(2): 1-7.
Yang Z H, Lu Z Y, Liu H. Online learning photovoltaic power prediction based on leakage integral echo state network［J］. Journal of Power System and Automation, 2018, 30(2): 1-7.
［10］Antonelo E A, Camponogara E, Plucenio A. System Identification of a Vertical Riser Model with Echo State Networks［J］.IFAC PapersOnLine,2015,48(6):304-310.
［11］伦淑娴, 胡海峰. 基于罚函数内点法的泄露积分型回声状态网的参数优化［J］. 自动化学报, 2017, 43(7): 1160-1168.
Lun S X, Hu H F. Parameter Optimization of Leakage Integral Echo State Network Based on Penalty Function Interior Point Method［J］. Journal of Automation, 2017, 43(7): 1160-1168.
［12］Jaeger H. Survey: Reservoir computing approaches to recurrent neural network training［M］. Netherlands: Elsevier Science Publishers, 2009.
［13］韩敏, 许美玲, 王新迎. 多元时间序列的子空间回声状态网络预测模型［J］. 计算机学报, 2014, 37(11): 2268-2275.
Han M, Xu M L, Wang X Y. Subspace Echo State Network Prediction Model for Multivariate Time Series［J］. Journal of Computer Science, 2014, 37(11): 2268-2275.
［14］宋彤, 李菡. 基于小波回声状态网络的混沌时间序列预测［J］. 物理学报, 2012, 61(8): 90-96.
Song T, Li H. Prediction of chaotic time series based on wavelet echo state network［J］. Chinese Journal of Physics, 2012, 61(8): 90-96.
［15］杨理践, 邢燕好, 张佳. 基于电磁超声导波的铝板裂纹缺陷检测方法［J］. 仪器仪表学报, 2018, 39(4): 150-160.
Yang L J, Xing Y H, Zhang J. Crack detection method for aluminum plate based on electromagnetic ultrasonic guided wave［J］. Journal of Instruments and Instruments, 2018, 39(4): 150-160.
［16］Song X C. Beijing PM2.5 DATA［DB］. http://archive.ics.uci.edu/ml, 2018-01-19.
［17］Li R, Hardy R, Zhang W, et al. Chemical Characterization and Source Apportionment of PM2. 5 in a Nonattainment Rocky Mountain Valley［J］. Journal of Environmental Quality, 2018, 47(2): 238-238.
［18］Aldabe J, Elustondo D, Santamaria C, et al. Chemical characterisation and source apportionment of PM2.5 and PM10 at rural, urban and traffic sites in Navarra (North of Spain)［J］.Atmospheric Research,2011,102(1):191-205.