Prediction of fine particulate matter concentrations based on generalized hidden Markov model

doi:10.11949/j.issn.0438-1157.20171113

Abstract

Abstract:

In recent years, severe haze pollution accidents occurred frequently, which caused heavy losses in national economy and health of residents in China. Accurate early warning of severe haze pollution episodes can not only remind people of refraining from the hazards, but also set aside enough time for the government of emergency management before substantial improvement of air quality. According to the non-Gaussian distribution characteristics of PM_2.5 precursors and meteorological factors and limitation of the number of known hidden states in traditional hidden Markov models (HMMs), generalized hidden Markov models (GHMMs) were employed to make predictions of PM_2.5 concentrations at 11 nationally controlled monitoring sites (except Dingling site) in Beijing from January 2013 to January 2017. Data from Jan. 2013 to Dec. 2015 was used to train the GHMM models and data from Jan. 2016 to Jan. 2017 was used to validate these models. The same data was also used to train traditional HMM which contained 2 hidden states and 6 Gaussian distributions to make comparison with GHMM. Results illustrate that true prediction rate of GHMMs is significantly higher than that of traditional HMMs when applied on the prediction of samples over 250 μg·m⁻³; while both GHMMs and HMMs have similar performances when applied on the prediction of samples blow 150 μg·m⁻³.

Key words: systems engineering, environment, pollution, fine particulate matter, prediction, algorithm, generalized hidden Markov model

摘要：

近年来，以PM_2.5为主要污染物的重霾污染事件频频发生，给我国国民经济和居民健康造成了严重损失。在空气质量尚未得到根本性改善的情况下，对重霾污染的准确预警不仅能使公众合理回避污染危害，还能为政府实施应急管理提供时间裕量。针对影响PM_2.5浓度的前体物及气象因素的非高斯分布特点以及传统隐马尔可夫模型（hidden Markov model，HMM）必须已知隐含状态个数的缺点，采用广义隐马尔可夫模型（generalized hidden Markov model，GHMM）对北京市除去定陵外的11个国控站点2013年1月～2017年1月的PM_2.5浓度进行了预测。结果表明：GHMM对严重污染及以上PM_2.5样本浓度预测准确率显著高于传统连续HMM，但针对中度污染及以下PM_2.5样本浓度的预测准确率接近传统HMM。

关键词: 系统工程, 环境, 污染, PM_2.5, 预测, 算法, 广义隐马尔可夫模型

CLC Number:

X513

ZHANG Hao, YU Junyi, LIU Xiaohui, LEI Hong. Prediction of fine particulate matter concentrations based on generalized hidden Markov model[J]. CIESC Journal, 2018, 69(3): 1215-1220.

张浩, 于君毅, 刘晓慧, 雷洪. 基于广义隐马尔可夫模型的PM_2.5浓度预测[J]. 化工学报, 2018, 69(3): 1215-1220.

References 30

[1]	KLEEMAN M J, SCHAUER J J, CASS G R. Size and composition distribution of fine particulate matter emitted from motor vehicles[J]. Environmental Science & Technology, 2000, 34(7):1132-1142.
[2]	温昶, 徐明厚, 于敦喜, 等. 煤粉O2/CO2燃烧时PM2.5及其Fe, S的生成特性[J]. 化工学报, 2011, 62(4):1062-1069. WEN C, XU M H, YU D X, et al. Formation characteristics of PM2.5 including Fe, S during O2/CO2 combustion of pulverized coal[J]. CIESC Journal, 2011, 62(4):1062-1069.
[3]	ZHANG Q, QUAN J N, TIE X X, et al. Effects of meteorology and secondary particle formation on visibility during heavy haze events in Beijing, China[J]. Science of the Total Environment, 2015, 502:578-584.
[4]	尹文君, 张大伟, 严京海, 等. 基于深度学习的大数据空气污染预报[J]. 中国环境管理, 2015, (6):46-52. YIN W J, ZHANG D W, YAN J H, et al. Deep learning based air pollutant forecasting with big data[J]. Chinese Journal of Environmental Management, 2015, (6):46-52.
[5]	MA Z W, HU X F, HUANG L, et al. Estimating ground-level PM2.5 in China using satellite remote sensing[J]. Environmental Science & Technology, 2014, 48(13):7436-7444.
[6]	HU J, ZHANG H, CHEN S H, et al. Predicting primary PM2.5 and PM0.1 trace composition for epidemiological studies in California[J]. Environmental Science & Technology, 2014, 48(9):4971-4979.
[7]	MAKAR P A, GONG W, MILBRANDT J, et al. Feedbacks between air pollution and weather (Ⅰ):Effects on weather[J]. Atmospheric Environment, 2015, 115:442-469.
[8]	MAKAR P, GONG W, HOGREFE C, et al. Feedbacks between air pollution and weather(Ⅱ):Effects on chemistry[J]. Atmospheric Environment, 2015, 115:499-526.
[9]	COBOURN W G. An enhanced PM2.5 air quality forecast model based on nonlinear regression and back-trajectory concentrations[J]. Atmospheric Environment, 2010, 44(25):3015-3023.
[10]	SUN W, ZHANG H, PALAZOGLU A, et al. Prediction of 24-hour-average PM2.5 concentrations using a hidden Markov model with different emission distributions in Northern California[J]. Science of the Total Environment, 2013, 443:93-103.
[11]	ZHANG H, ZHANG W D, PALAZOGLU A, et al. Prediction of ozone levels using a hidden Markov model (HMM) with gamma distribution[J]. Atmospheric Environment, 2012, 62:64-73.
[12]	ZHOU G Q, XU J M, XIE Y, et al. Numerical air quality forecasting over eastern China:an operational application of WRF-Chem[J]. Atmospheric Environment, 2017, 153:94-108.
[13]	REISEN V A, SARNAGLIA A J Q, REIS JR N C, et al. Modeling and forecasting daily average PM10 concentrations by a seasonal long-memory model with volatility[J]. Environmental Modelling & Software, 2014, 51:286-295.
[14]	苏文, 赵力, 邓帅. 基于基团拓扑的遗传神经网络工质临界温度预测[J]. 化工学报, 2016, 67(11):4689-4695. SU W, ZHAO L, DENG S. Prediction of refrigerant critical temperature with genetic neural network based on group topology[J]. CIESC Journal, 2016, 67(11):4689-4695.
[15]	黄凯, 陈勇, 母志为, 等. 基于人工神经网络和遗传算法的甲烷制氢催化剂设计[J]. 化工学报, 2016, 67(8):3481-3490. HUANG K, CHEN Y, MU Z W, et al. Catalyst design for production of hydrogen from methane based on artificial neural network and genetic algorithm[J]. CIESC Journal, 2016, 67(8):3481-3490.
[16]	SUN W, ZHANG H, PALAZOGLU A. Prediction of 8 h-average ozone concentration using a supervised hidden Markov model combined with generalized linear models[J]. Atmospheric Environment, 2013, 81:199-208.
[17]	ROPERO R, AGUILERA P, FERNáNDEZ A, et al. Regression using hybrid Bayesian networks:modelling landscape-socioeconomy relationships[J]. Environmental Modelling & Software, 2014, 57:127-137.
[18]	LÜ B L, COBOURN W G, BAI Y Q. Development of nonlinear empirical models to forecast daily PM2.5 and ozone levels in three large Chinese cities[J]. Atmospheric Environment, 2016, 147:209-223.
[19]	RABINER L R. A tutorial on hidden Markov models and selected applications in speech recognition[J]. Proceedings of the IEEE, 1989, 77(2):257-286.
[20]	DEBYECHE M, HATON J P, HOUACINE A. A new vector quantization approach for discrete HMM speech recognition system[J]. International Journal of Computing, 2014, 5(1):72-78.
[21]	DIAS J G, VERMUNT J K, RAMOS S. Clustering financial time series:new insights from an extended hidden Markov model[J]. European Journal of Operational Research, 2015, 243(3):852-864.
[22]	KROGH A, LARSSON B, VON HEIJNE G, et al. Predicting transmembrane protein topology with a hidden Markov model:application to complete genomes[J]. Journal of Molecular Biology, 2001, 305(3):567-580.
[23]	YAMATO J, OHYA J, ISHⅡ K. Recognizing human action in time-sequential images using hidden markov model[C]//Proceedings of the Computer Vision and Pattern Recognition, 1992 IEEE Computer Society Conference on, 1992.
[24]	DONG M, YANG D, KUANG Y, et al. PM2.5 concentration prediction using hidden semi-Markov model-based times series data mining[J]. Expert Systems with Applications, 2009, 36(5):9046-9055.
[25]	ZHANG H, PALAZOGLU A, ZHANG X, et al. Prediction of surface ozone exceedance days using PCA with a non-parametric T2 control limit[J]. Chemometrics and Intelligent Laboratory Systems, 2014, 133:42-48.
[26]	SUN W, PALAZOGLU A, SINGH A, et al. Prediction of surface ozone episodes using clusters based generalized linear mixed effects models in Houston-Galveston-Brazoria area, Texas.[J]. Atmospheric Pollution Research, 2015, 6(2):245-253.
[27]	PFUNDSTEIN G. Hidden Markov models with generalised emission distribution for the analysis of high-dimensional, non-euclidean data[D]. Munich:Institut für Statistik, 2011.
[28]	SCHNEIDER T. Analysis of incomplete climate data:estimation of mean values and covariance matrices and imputation of missing values[J]. Journal of Climate, 2001, 14(5):853-871.
[29]	WANG Y L, BARBACIORU C, HYLAND F, et al. Large scale real-time PCR validation on gene expression measurements from two commercial long-oligonucleotide microarrays[J]. BMC Genomics, 2006, 7(1):59.
[30]	DUTOT A-L, RYNKIEWICZ J, STEINER F E, et al. A 24-h forecast of ozone peaks and exceedance levels using neural classifiers and weather predictions[J]. Environmental Modelling & Software, 2007, 22(9):1261-1269.