基于GOSSA和HMM的时间序列预测算法

doi:10.12263/DZXB.20220856

电子学报

• •

基于GOSSA和HMM的时间序列预测算法

李大社¹, 孙元威¹, 阮俊虎²

^1.山东工商学院计算机科学与技术学院，山东烟台 264000
^2.西北农林科技大学经济管理学院，陕西咸阳 712000

收稿日期:2022-07-19 修回日期:2022-08-27 出版日期:2023-03-07
- 作者简介:
- 李大社男，1978年2月出生于山东临沂，工学博士，现为山东工商学院计算机科学与技术学院副教授，研究方向为智能计算、图像处理等. E-mail： lidashe@126.com
  孙元威男，1998年5月出生于山东烟台.现为山东工商学院计算机科学与技术学院硕士研究生，主要研究方向为智能计算. E-mail：1609456079@qq.com
  阮俊虎男，1983年10月生于河南周口，工学博士，现为西北农林科技大学经济管理学院教授，研究方向为数字农业、农业大数据与互联网等. E-mail： rjh@nwsuaf.edu.cn
- 基金资助:
- 国家重点研发计划项目(2020YFD000204);国家自然科学基金(71973106);山东省重点研发计划(2020RKB01555);烟台市科技创新发展计划项目(2021XDHZ062)

Time Series Prediction Algorithm Based on GOSSA and HMM

LI Da-she¹, SUN Yuan-wei¹, RUAN Jun-hu²

^1.School of Computer Science and Technology，Shandong Technology and Business University，Yantai，Shandong 264000，China
^2.School of Economics and Management，Northwest Agriculture & Forestry University，XianYang，Shaanxi 712000，China

Received:2022-07-19 Revised:2022-08-27 Online:2023-03-07
- Supported by:
- China's National Key R&D Program(2020YFD000204);National Natural Science Foundation of China(71973106);Key R&D Program of Shandong Province, China(2020RKB01555);Yantai Science and Technology Innovation Development Project(2021XDHZ062)

摘要/Abstract

摘要：

时间序列具有非线性和不稳定性等特点，当前时间序列预测研究面临模型训练参数多、泛化能力差等挑战，其预测精度无法保证.基于此，本文提出一种基于全局最优的麻雀搜索算法（Globally Optimal Sparrow Search Algorithm，GOSSA）和隐马尔可夫模型（Hidden Markov Model， HMM）相融合的时间序列预测模型（GOSSA-HMM）.根据隐马尔可夫模型在模式识别和分类上的优势，对原始数据做差值处理并划分类别属性，以此作为隐马尔可夫模型的输入.采用全局最优的麻雀搜索算法对隐马尔可夫模型的参数进行训练，以解决参数训练过程中存在的收敛速度慢，对初始值设置敏感的问题.将赋予类别属性的差值数据进行分段，利用改进之后的隐马尔可夫模型测算每段序列走势的概率，从与当前数据走势相匹配的过去数据集中定位相同的模型实现预测.通过对山东半岛15个海洋牧场的溶解氧数据进行预测分析，结果表明与当前主要时间序列预测算法相比，GOSSA-HMM训练的参数较少，计算成本较低，具有更好的预测精度和泛化能力.

关键词: 时间序列预测, 隐马尔科夫模型, 麻雀搜索算法

Abstract:

Time series has the characteristics of non-linearity and instability. There exist some deficiencies in the current researches of time series prediction, such as too many training parameters and poor generalization, which leads to its low prediction accuracy. In order to solve the problems, this paper proposes GOSSA-HMM, a prediction model for time series, based on the fusion of global optimal sparrow search algorithm (GOSSA) and hidden markov model (HMM). By using the advantages of the hidden markov model in pattern recognition and classification, the original data in time series can be dealt with by the subtractive preprocessing and classification, which is used as the input of the Hidden Markov Model. The GOSSA is used to train the parameters of the HMM to solve the problems, such as slow convergence speed and being sensitive to the initial value setting. The D-value data endowed with category attributes are to be segmented, the probability of the trend of each sequence will be calculated by using the improved HMM, and the prediction can be achieved by locating the same pattern with the past datasets matching the trend of the current data. The dissolved oxygen data from 15 marine ranches in the Shandong Peninsula are combined for analysis. Under the same experimental conditions, GOSSA-HMM has fewer training parameters, a lower cost of calculation, and better prediction accuracy and generalization.

Key words: time series, hidden markov model, sparrow search algorithm, prediction

中图分类号:

TP391

李大社, 孙元威, 阮俊虎. 基于GOSSA和HMM的时间序列预测算法[J]. 电子学报, DOI: 10.12263/DZXB.20220856.

Da-she LI, Yuan-wei SUN, Jun-hu RUAN . Time Series Prediction Algorithm Based on GOSSA and HMM[J]. Acta Electronica Sinica, DOI: 10.12263/DZXB.20220856.

图/表 11

图1 隐马尔可夫模型的基本过程

图2 迭代次数与Sin混沌映射关系(a) 映射函数值折线图 (b) 映射函数值散点图

图3 发现者位置更新改进前后对比(a) f(x)=exp(-x/α·Maxiter) (b) f(x)=1+Z

图4 GOSSA-HMM预测模型流程

图5 参数敏感性分析热力图

表1 不同预测模型误差比较

	GOSSA-HMM	ISSA-HMM	HMM	LSTM	TCN	Transformer
MAE	0.0419	0.0470	0.0778	0.1142	0.0719	0.0943
MSE	0.0060	0.0072	0.0160	0.0349	0.0125	0.0217
RMSE	0.0777	0.0851	0.1266	0.1870	0.1121	0.1475
MAPE	0.0058	0.0067	0.0110	0.0165	0.0101	0.0133
R²	0.9950	0.9929	0.9863	0.9694	0.9885	0.9814

图6 不同算法模型的真实值与预测值的比较

表2 不同预测模型复杂度和效率对比

O(TN²)+

O(nd)

O(TN²)+

O(nd)

图7 15个海洋牧场真实值与预测值比较

图8 预测模型性能的泰勒图

图9 预测模型性能的雷达图

参考文献 38

1	TAPIA S, KRISTJANPOLLER W. Framework based on multiplicative error and residual analysis to forecast bitcoin intraday-volatility[J]. Physica A: Statistical Mechanics and its Applications, 2022, 589: 126613.
2	GREENAWAY-MCGREVY R. Forecast combination for VARs in large N and T panels[J]. International Journal of Forecasting, 2022, 38(1): 142-164.
3	YAN H, QI Y, YU D J. Short-term traffic flow prediction based on a hybrid optimization algorithm[J]. Applied Mathematical Modelling, 2022, 102: 385-404.
4	冯宁, 郭晟楠, 宋超, 朱琪超, 万怀宇. 面向交通流量预测的多组件时空图卷积网络[J]. 软件学报, 2019, 30(03): 759-769.
	FENG Ning, GUO Sheng-nan, SONG Chao, ZHU Qi-chao, WAN Hai-yu. Multi-component spatial-temporal graph convolution networks for traffic flow forecasting[J]. Journal of Software, 2019, 30(3): 759-769. (in Chinese)
5	张兴辉, 樊秀梅, 阿喜达, 樊书嘉, 武文瑜. 反向学习的灰狼算法优化及其在交通流预测中的应用[J]. 电子学报, 2021, 49(05): 879-886.
	ZHANG Xing-hui, FAN Xiu-mei, Xi-da E, FAN Shu-jia, WU Wen-yu. Grey Wolf Optimization Based on Opposition Learning and Its Application in Traffic Flow Forecasting[J]. Acta Electronica Sinica, 2021, 49(05): 879-886. (in Chinese)
6	蒋锋, 何佳琪, 曾志刚, 田天海. 基于分解–优化–集成学习方法的电价预测[J]. 中国科学:信息科学, 2018,48(10): 1300-1315.
	JIANG Feng, HE Jia-qi, ZENG Zhi-gang, TIAN Tian-hai. Decomposition-optimization-ensemble learning approach for electricity price forecasting[J]. SCIENTIA SINICA Informationis, 2018, 48(10): 1300-1315. (in Chinese)
7	郑威迪, 李志刚, 贾涵中, 高闯. 基于改进型鲸鱼优化算法和最小二乘支持向量机的炼钢终点预测模型研究[J]. 电子学报, 2019, 47(03): 700-706.
	ZHENG Wei-di, LI Zhi-gang, JIA Han-zhong, GAO Chuang. Research on Prediction Model of Steelmaking End Point Based on LWOA and LSSVM[J]. Acta Electronica Sinica, 2019, 47(03): 700-706. (in Chinese)
8	熊有成, 赵鸿. 长短期记忆网络预测混沌时间序列[J]. 中国科学:物理学力学天文学, 2019, 49(12): 92-99.
	XIONG You-cheng, ZHAO Hong. Chaotic time series prediction based on long short-term memory neural networks[J]. SCIENTIA SINICA Physica, Mechanica & Astronomica, 2019, 49(12): 92-99. (in Chinese)
9	KESHTEGAR B, HEDDAM S. Modeling daily dissolved oxygen concentration using modified response surface method and artificial neural network: a comparative study[J]. Neural Computing and Applications, 2018, 30(10): 2995-3006.
10	Tao Y, Wang Y, Wang D, et al. A C-vine copula framework to predict daily water temperature in the Yangtze River[J]. Journal of Hydrology, 2021, 598: 126430.
11	Yang A, Li W, Yang X. Short-term electricity load forecasting based on feature selection and Least Squares Support Vector Machines[J]. Knowledge-Based Systems, 2019, 163: 159-173.
12	SOHRABI M M, BeENANKAR R, ToOINA D, et al. Estimation of daily stream water temperatures with a Bayesian regression approach[J]. Hydrological Processes, 2017, 31(9): 1719-1733.
13	KISI O, SHIRI J, KARIMI S, et al. A survey of water level fluctuation predicting in Urmia Lake using support vector machine with firefly algorithm[J]. Applied Mathematics and Computation, 2015, 270: 731-743.
14	OLYAIE E, ABYANEH H Z, MEHR A D. A comparative analysis among computational intelligence techniques for dissolved oxygen prediction in Delaware River[J]. Geoscience Frontiers, 2017, 8(3): 517-527.
15	金苍宏, 董腾然, 陈天翼, 吴明晖, 李国军, 周胜利. 融合序列分解与时空卷积的时序预测算法[J]. 电子学报, 2021, 49(02): 233-238.
	JIN Cang-hong, DONG Teng-ran, CHEN Tian-yi, WU Ming-hui, LI Guo-jun, ZHOU Sheng-li. Spatio-Temporal Convolutional Forecasting Based on Time-Series Decomposition Strategy[J]. Acta Electronica Sinica, 2021, 49(02): 233-238. (in Chinese)
16	Yu P, Yan X. Stock price prediction based on deep neural networks[J]. Neural Computing and Applications, 2020, 32(6): 1609-1628.
17	Wan H, Guo S, Yin K, et al. CTS-LSTM: LSTM-based neural networks for correlated time series prediction[J]. Knowledge-Based Systems, 2020, 191: 105239.
18	SAGHEER A, KOTB M. Time series forecasting of petroleum production using deep LSTM recurrent networks[J]. Neurocomputing, 2019, 323: 203-213.
19	JAHANBAKHT M, Xiang W, AZGHADI M R. Sediment Prediction in the Great Barrier Reef using Vision Transformer with finite element analysis[J]. Neural Networks, 2022, 152: 311-321.
20	谢贵才, 段磊, 蒋为鹏, 肖珊, 徐一凡. 多尺度时序依赖的校园公共区域人流量预测[J]. 软件学报, 2021,32(03): 831-844.
	XIE Gui-cai. DUAN Lei, JIANG Wei-peng, XIAO Shan, XU Yi-fan. Pedestrian volume prediction for campus public area based on multi-scale temporal dependency[J]. Journal of Software, 2021, 32(03): 831-844. (in Chinese)
21	Tan C, Liu X, Zhang G. Inferring Brain State Dynamics Underlying Naturalistic Stimuli Evoked Emotion Changes With dHA-HMM[J]. Neuroinformatics, 2022: 1-17.
22	Gao R, Yu J, Zhang M, et al. A mathematical formulation of the central dogma of molecular biology[M]//Nanomedicine. Jenny Stanford Publishing, 2019: 81-116.
23	Han W, Wu Y, Zeng L, et al. Building the Chordata Olfactory Receptor Database using more than 400,000 receptors annotated by Genome2OR[J]. Science China Life Sciences, 2022: 1-13.
24	陈哲怀, 郑文露, 游永彬, 钱彦旻, 俞凯. 标签同步解码算法及其在语音识别中的应用[J]. 计算机学报, 2019, 42(07): 1511-1523.
	CHEN Zhe-hua, ZHENG Wen-lou, YOU Yong-bin, QIAN Yan-min, YU Kai. Label Synchronous Decoding for Speech Recognition[J]. Journal of Computers, 2019,42(07): 1511-1523. (in Chinese)
25	龙华, 杨明亮, 邵玉斌. 基于特征流融合的带噪语音检测算法[J]. 通信学报, 2020, 41(04): 134-142.
	LONG Hua, YANG Ming-liang, SHAO Yu-bin. Noisy voice detection algorithm based on feature stream fusion[J]. Journal on Communications, 2020, 41(04): 134-142. (in Chinese)
26	Zhang M, Jiang X, Fang Z, et al. High-order Hidden Markov Model for trend prediction in financial time series[J]. Physica A: Statistical Mechanics and its Applications, 2018, 517: 1-12.
27	Liu Y, Ye L, Qin H, et al. Monthly streamflow forecasting based on hidden Markov model and Gaussian Mixture Regression[J]. Journal of Hydrology, 2018, 561:146-159.
28	Xiong X, Chen L, Liang J. A new framework of vehicle collision prediction by combining SVM and HMM[J]. IEEE Transactions on Intelligent Transportation Systems, 2017, 19(3): 699-710.
29	Liao W, Li D, Cui S. A heuristic optimization algorithm for HMM based on SA and EM in machinery diagnosis[J]. Journal of Intelligent Manufacturing, 2018, 29(8): 1845-1857.
30	TENYAKOV A, MAMON R, DAVISON M. Modelling high-frequency FX rate dynamics: A zero-delay multi-dimensional HMM-based approach[J]. Knowledge-Based Systems, 2016, 101: 142-155.
31	SAGAYAM K M, HEMANTH D J. ABC algorithm based optimization of 1-D hidden Markov model for hand gesture recognition applications[J]. Computers in Industry, 2018, 99: 313-323.
32	Zhang C, Ding S. A stochastic configuration network based on chaotic sparrow search algorithm[J]. Knowledge-Based Systems, 2021, 220: 1-20.
33	李爱莲, 全凌翔, 崔桂梅, 解韶峰. 融合正余弦和柯西变异的麻雀搜索算法[J]. 计算机工程与应用, 2022, 58(03): 91-99.
	LI Ai-qin, QUAN Ling-xiang, CUI Gui-mei, XIE Shao-feng. Sparrow Search Algorithm Combining Sine-Cosine and Cauchy Mutation[J]. Computer Engineering and Applications, 2022, 58(03): 91-99. (in Chinese)
34	张晓萌, 张艳珠, 刘禄, 张硕, 熊夫睿. 融合多策略的改进麻雀搜索算法[J]. 计算机应用研究, 2022, 39(04): 1086-1091+1117.
	ZHANG Xiao-meng, ZHANG Yan-zhu, LIU Lu, ZHANG Shuo, XIONG Fu-rui. Improved sparrow search algorithm fused with multiple strategies[J]. Application Research of Computers, 2022,39(04): 1086-1091+1117. (in Chinese)
35	王永, 赵毅, Gao Jerry, 陈燕. 基于分段Logistic映射的二维耦合映像格子模型的密码学相关特性分析[J]. 电子学报, 2019, 47(03): 657-663.
	WANG Yong, ZHAO Yi, Gao Jerry, CHEN Yan. Cryptographic Feature Analysis on 2D Coupled Map Lattices Based on Piecewise Logistic Map[J]. Acta Electronica Sinica, 2019,47(03): 657-663. (in Chinese)
36	Yang Y, Zhang H, Yan P, et al. Multi-objective optimization for efficient modeling and improvement of the high temperature PEM fuel cell based Micro-CHP system[J]. International Journal of Hydrogen Energy, 2020, 45(11): 6970-6981.
37	杨海东, 鄂加强. 自适应变尺度混沌免疫优化算法及其应用[J]. 控制理论与应用, 2009, 26(10): 1069-1074.
	YANG Hai-dong, Jia-qiang E. An adaptive chaos immune optimization algorithm with mutative scale and its application[J]. Control Theory & Applications, 2009, 26(10): 1069-1074. (in Chinese)
38	KARGAR K, SAMADIANFARD S, PARSA J, et al. Estimating longitudinal dispersion coefficient in natural streams using empirical models and machine learning algorithms[J]. Engineering Applications of Computational Fluid Mechanics, 2020, 14(1): 311-322.

[1]	徐佳伟, 罗倩. 基于遗传非参数MDL-BW方法的HMM结构优化[J]. 电子学报, 2022, 50(11): 2765-2772.
[2]	金苍宏, 董腾然, 陈天翼, 吴明晖, 李国军, 周胜利. 融合序列分解与时空卷积的时序预测算法[J]. 电子学报, 2021, 49(2): 233-238.
[3]	雷涛, 张肖, 加小红, 刘侍刚, 张艳宁. 基于模糊聚类的图像分割研究进展[J]. 电子学报, 2019, 47(8): 1776-1791.
[4]	房丙午, 黄志球, 王勇, 李勇. 状态不可观测的信息物理融合系统运行时验证[J]. 电子学报, 2018, 46(12): 2824-2831.
[5]	汪成亮, 王小均. 基于三轴传感器的老年人日常活动识别[J]. 电子学报, 2017, 45(3): 570-576.
[6]	章登义, 欧阳黜霏, 吴文李. 基于N阶近邻路网的车辆行程时间估计模型[J]. 电子学报, 2015, 43(12): 2491-2496.
[7]	章登义, 欧阳黜霏, 吴文李. 针对时间序列多步预测的聚类隐马尔科夫模型[J]. 电子学报, 2014, 42(12): 2359-2364.
[8]	徐小来;雷英杰;谢文彪. 基于UKF的自组织直觉模糊神经网络[J]. 电子学报, 2010, 38(3): 638-645.
[9]	刘大同, 彭宇, 彭喜元. 基于残差预测修正的局部在线时间序列预测方法[J]. 电子学报, 2008, 36(S1): 81-85.
[10]	彭喜元;王军;彭宇. 一种新型时间序列多分辨预测模型研究[J]. 电子学报, 2007, 35(11): 2146-2149.
[11]	刘芳, 刘文学, 焦李成. 基于复小波邻域隐马尔科夫模型的图像去噪[J]. 电子学报, 2005, 33(7): 1284-1287.
[12]	左国玉;刘文举;阮晓钢. 声音转换技术的研究与进展[J]. 电子学报, 2004, 32(7): 1165-1172.

基于GOSSA和HMM的时间序列预测算法

Time Series Prediction Algorithm Based on GOSSA and HMM

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 38

相关文章 12

编辑推荐

Metrics