A Comparison of Nonlinear Regression and Neural Network Models for Ground-Level Ozone Forecasting

ABSTRACT

A hybrid nonlinear regression (NLR) model and a neural network (NN) model, each designed to forecast next-day maximum 1-hr average ground-level O₃ concentrations in Louisville, KY, were compared for two O₃ seasons—1998 and 1999. The model predictions were compared for the forecast mode, using forecasted meteorological data as input, and for the hindcast mode, using observed meteorological data as input. The two models performed nearly the same in the forecast mode. For the two seasons combined, the mean absolute forecast error was 12.5 ppb for the NLR model and 12.3 ppb for the NN model. The detection rate of 120 ppb threshold exceedances was 42% for each model in the forecast mode. In the hindcast mode, the NLR model performed marginally better than the NN     

Neural Network and Multiple Regression Models for PM10 Prediction in Athens: A Comparative Assessment

Abstract

Particulate atmospheric pollution in urban areas is considered to have significant impact on human health. Therefore, the ability to make accurate predictions of particulate ambient concentrations is important to improve public awareness and air quality management. This study examines the possibility of using neural network methods as tools for daily average particulate matter with aerodynamic diameter <10 µm (PM₁₀) concentration forecasting, providing an alternative to statistical models widely used up to this day. Based on a data inventory, in a fixed central site in Athens, Greece, ranging over a two-year period, and using mainly meteorological variables as inputs, neural network models and multiple linear regression models were developed and evaluated. Comparison statistics used indicate that the neural network approach has an edge over regression models, expressed both in terms of prediction error (root mean square error values lower by 8.2–9.4%) and of episodic prediction ability (false alarm rate values lower by 7–13%). The results demonstrate that artificial neural networks (ANNs), if properly trained and formed, can provide adequate solutions to particulate pollution prognostic demands.     

基于人工神经网络和遗传算法研究锰过氧化酶脱色甲基橙

基于锰过氧化物酶(MnP)氧化脱色偶氮类染料的原理,实验研究MnP对甲基橙的脱色工艺,采用人工神经网络(ANN)和遗传算法(GA)建立脱色模型并优化工艺。建立的ANN模型的误差、相关系数、均方根误差和绝对平均偏差分别为0.0009、0.9971、1.21和6.82,模型有效且能够用于预测和工艺优化。采用GA对ANN模型进行数值寻优,得到的最佳工艺条件为酶液量0.6 mL,Mn2+浓度4 mmol/L,H2O2浓度0.49 mmol/L。该条件下脱色率达到(90.74±0.59)%。ANN耦合GA有效地建立了锰过氧化物酶脱色甲基橙的模型,并优化了工艺参数,为甲基橙脱色的研究提供一定参考。     

Prediction of NO and NO2 concentrations near a street with heavy traffic in Santiago,Chile

Based on NO concentrations and meteorological variables recorded hourly at a point close to an avenue with heavy traffic in the city of Santiago, we are able to build a simple model that allows prediction of NO concentrations several hours in advance. Predicted NO concentrations in conjunction with forecasted meteorological data may be used to predict NO₂ concentrations with reasonable accuracy. We compare predictions generated using persistence, linear regressions and multi layer neural networks.     

Prediction of sulfur dioxide concentrations at a site near downtown Santiago,Chile

We have analyzed the possibility to predict hourly averages of sulfur dioxide concentrations in the atmosphere at a site not far from the downtown area in the city of Santiago, Chile. We have compared the forecasts produced assuming persistence, linear regressions and feed forward neural networks. The effect of meteorological conditions is included by using forecasted values of temperature, relative humidity and wind speed at the time of the intended prediction as inputs to the different models. The best predictions for hourly averages are obtained with a three-layer neural network that has hourly averages of sulfur dioxide concentrations every 6 h on the previous day plus the actual values of the meteorological variables as input. Training the network with 1995 data, error in 8 h in advance prediction for 1996 data is of the order of 30%.     

Neural network and multiple regression models for PM10 prediction in Athens: a comparative assessment

Particulate atmospheric pollution in urban areas is considered to have significant impact on human health. Therefore, the ability to make accurate predictions of particulate ambient concentrations is important to improve public awareness and air quality management. This study examines the possibility of using neural network methods as tools for daily average particulate matter with aerodynamic diameter <10 microm (PM10) concentration forecasting, providing an alternative to statistical models widely used up to this day. Based on a data inventory, in a fixed central site in Athens, Greece, ranging over a two-year period, and using mainly meteorological variables as inputs, neural network models and multiple linear regression models were developed and evaluated. Comparison statistics used indicate that the neural network approach has an edge over regression models, expressed both in terms of prediction error (root mean square error values lower by 8.2-9.4%) and of episodic prediction ability (false alarm rate values lower by 7-13%). The results demonstrate that artificial neural networks (ANNs), if properly trained and formed, can provide adequate solutions to particulate pollution prognostic demands.     

应用神经网络方法优化垃圾焚烧模拟灰渣熔点的预测

采用SiO₂、Al₂O₃、CaO、Na₂CO₃、NaCl和Fe₂O₃等物质来模拟垃圾焚烧的真实灰渣组成,通过实验测定模拟灰渣熔点,建立神经网络模型进行熔点预测,由预测结果来指导进一步实验,得到修正的模型,最终预测出的半球温度（HT）平均误差低于5%。     

生物脱氮除磷活性污泥系统复合模拟方法

为避免繁琐的参数校核工作,提出了活性污泥2 d号模型(ASM2d)和人工神经网络(ANNs)相结合的复合模拟方法。考察了复合方法在某污水处理厂生物脱氮除磷工艺中的应用情况。研究表明,ANNs能够准确地模拟出水实测值与未经校核的ASM2d机理模型的估计值之间的差值。利用Levenberg-Marquardt算法,对出水氨氮、总氮和总磷分别建立网络结构为5-12-1、5-8-1和5-8-1的ANNs子模型,将这些子模型输出同ASM2d机理模型输出相加便得到复合模型输出。复合模型估计值对前10.4 d(ANNs子模型训练数据时段)出水氨氮、总氮和总磷浓度的拟合平均绝对百分比误差分别为0.267、0.055和0.048;其对后2.6 d(ANNs子模型测试数据时段)出水氨氮、总氮和总磷浓度的预测平均绝对百分比误差分别为0.332、0.083和0.069。均方根误差、平均绝对误差等评价指标也表明复合模型能够给出合理的模拟结果。     

Prediction of maximum of 24-h average of PM10 concentrations 30 h in advance in Santiago,Chile

We have developed a neural network based model that uses values of PM10 concentrations measured until 6 p.m. on the present day plus measured and forecasted values of meteorological variables as input in order to predict the level reached by the maximum of the 24-h moving average (24MA) of PM10 concentration on the next day. We have adjusted the parameters of the model using 1998 data to predict 1999 conditions and 1999 data to forecast 2000 maximum concentrations. We have found that among the relevant meteorological input variables, the forecasted difference between maximum and minimum temperature is the most important. Due to the fact that local authorities impose restrictions to emissions on days when the maximum of 24MA of PM10 concentration is expected to exceed 240 μg/m³, we have corrected the measured concentrations on these days before evaluating the efficacy of the forecasting model. Percent errors in forecasting the numerical value are of the order of 20%. The performance of the neural network is better than that of a linear model with the same inputs, but the difference being not great is an indication that the right choice of input variables may be more important than the decision to use a linear or a nonlinear model.     

Evaluation of PM10 forecasting based on the artificial neural network model and intake fraction in an urban area: A case study in Taiyuan City,China

Primary fine particulate matters with a diameter of less than 10 µm (PM₁₀) are important air emissions causing human health damage. PM₁₀ concentration forecast is important and necessary to perform in order to assess the impact of air on the health of living beings. To better understand the PM₁₀ pollution health risk in Taiyuan City, China, this paper forecasted the temporal and spatial distribution of PM₁₀ yearly average concentration, using Back Propagation Artificial Neural Network (BPANN) model with various air quality parameters. The predicted results of the models were consistent with the observations with a correlation coefficient of 0.72. The PM₁₀ yearly average concentrations combined with the population data from 2002 to 2008 were given into the Intake Fraction (IF) model to calculate the IFs, which are defined as the integrated incremental intake of a pollutant released from a source category or a region over all exposed individuals. The results in this study are only for main stationary sources of the research area, and the traffic sources have not been included. The computed IFs results are therefore under-estimations. The IFs of PM₁₀ from Taiyuan with a mean of 8.5 per million were relatively high compared with other IFs of the United States, Northern Europe and other cities in China. The results of this study indicate that the artificial neural network is an effective method for PM₁₀ pollution modeling, and the Intake Fraction model provides a rapid population risk estimate for pollutant emission reduction strategies and policies.

Implications The PM₁₀ (particulate matter with an aerodynamic diameter ≤10 μm) yearly average concentration of Taiyuan, with a mean of 0.176 mg/m³, was higher than the 65 μg/m³ recommended by the U.S. Environmental Protection Agency (EPA). The spatial distribution of PM₁₀ yearly average concentrations showed that wind direction and wind speed played an important role, whereas temperature and humidity had a lower effect than expected. Intake fraction estimates of Taiyuan were relatively high compared with those observed in other cities. Population density was the major factor influencing PM₁₀ spatial distribution. The results indicated that the artificial neural network was an effective method for PM₁₀ pollution modeling.     

Impacts of Meteorological Factors on PM10: Artificial Neural Networks (ANN) and Multiple Linear Regression (MLR) Approaches

In this study, prediction capacities of multi-linear regression (MLR) and artificial neural networks (ANN) onto coarse particulate matter (PM₁₀) concentrations were investigated. Different meteorological factors on particulate pollution were chosen for operating variables in the model analyses. Two different regions (urban and industrial) were identified in the region of Kocaeli, Turkey. All data sets were obtained from air quality monitoring network of the Ministry of Environment and Urban Planning, and 120 data sets were used in the MLR and ANN models. Regression equations explained the effects of the meteorological factors in MLR analyses. In the ANN model, backpropagation network with two hidden layers has achieved the best prediction efficiency. Determination coefficients and error values were examined for each model. ANN models displayed more accurate results compared to MLR.     

Neural networks and periodic components used in air quality forecasting

Forecasting of air quality parameters is one topic of air quality research today due to the health effects caused by airborne pollutants in urban areas. The work presented here aims at comparing two principally different neural network methods that have been considered as potential tools in that area and assessing them in relation to regression with periodic components. Self-organizing maps (SOM) represent a form of competitive learning in which a neural network learns the structure of the data. Multi-layer perceptrons (MLPs) have been shown to be able to learn complex relationships between input and output variables. In addition, the effect of removing periodic components is evaluated with respect to neural networks. The methods were evaluated using hourly time series of NO₂ and basic meteorological variables collected in the city of Stockholm in 1994–1998. The estimated values for forecasting were calculated in three ways: using the periodic components alone, applying neural network methods to the residual values after removing the periodic components, and applying only neural networks to the original data. The results showed that the best forecast estimates can be achieved by directly applying a MLP network to the original data, and thus, that a combination of the periodic regression method and neural algorithms does not give any advantage over a direct application of neural algorithms.     

Neural network modelling and prediction of hourly NOx and NO2 concentrations in urban air in London

Multilayer perceptron (MLP) neural networks were trained to model hourly NO_x and NO₂ pollutant concentrations in Central London from basic hourly meteorological data. Results have shown that the models perform well when compared to previous attempts to model the same pollutants using regression based models. This work also illustrates that MLP neural networks are capable of resolving complex patterns of source emissions without any explicit external guidance.     

Neural network approach for modelling ammonia emission after manure application on the field

This paper presents a neural network approach, which enables one to simulate ammonia emission after manure application on the field. Based on the data from 227 experiments out of previously published research, it can be illustrated that the time course of accumulated ammonia emission follows a non-linear Michaelis–Menten-like function. This function is determined by the two parameters E_max and K_M, which are dependent on manure-specific driving forces, application parameters and climate. 102 data sets of the 227 experiments showed sufficient data for training and validating neural networks for estimating E_max and K_M. The neural networks could be trained to R² values of 0.926 and 0.832 for the training set and the validation set of E_max, and to R² values of 0.988 for the training set and 0.527 for the validation set of the K_M-value, respectively.     

Prediction of dissolved oxygen concentration in hypoxic river systems using support vector machine: a case study of Wen-Rui Tang River,China

Accurate quantification of dissolved oxygen (DO) is critically important for managing water resources and controlling pollution. Artificial intelligence (AI) models have been successfully applied for modeling DO content in aquatic ecosystems with limited data. However, the efficacy of these AI models in predicting DO levels in the hypoxic river systems having multiple pollution sources and complicated pollutants behaviors is unclear. Given this dilemma, we developed a promising AI model, known as support vector machine (SVM), to predict the DO concentration in a hypoxic river in southeastern China. Four different calibration models, specifically, multiple linear regression, back propagation neural network, general regression neural network, and SVM, were established, and their prediction accuracy was systemically investigated and compared. A total of 11 hydro-chemical variables were used as model inputs. These variables were measured bimonthly at eight sampling sites along the rural-suburban-urban portion of Wen-Rui Tang River from 2004 to 2008. The performances of the established models were assessed through the mean square error (MSE), determination coefficient (R ²), and Nash-Sutcliffe (NS) model efficiency. The results indicated that the SVM model was superior to other models in predicting DO concentration in Wen-Rui Tang River. For SVM, the MSE, R ², and NS values for the testing subset were 0.9416 mg/L, 0.8646, and 0.8763, respectively. Sensitivity analysis showed that ammonium-nitrogen was the most significant input variable of the proposal SVM model. Overall, these results demonstrated that the proposed SVM model can efficiently predict water quality, especially for highly impaired and hypoxic river systems.     

BP神经网络对蚯蚓滤池处理COD的模拟预测

基于蚯蚓滤池处理去除污染物的非线性特点,利用BP神经网络建立了蚯蚓滤池处理COD的基本模型结构。同时对实验数据进行了验证和预测,通过权值贡献率分析确定了各种输入因素对COD出水浓度的影响。结果表明:COD的出水模型预测值与实际值平均误差较小,模型稳定,预测效果好。输入神经元为4,隐含神经元为8,输出神经元为1,学习速率为0.1,动量为0.1,训练次数为10 000的BP神经网络模型,预测的COD出水值最接近真实值。COD进水浓度对COD出水影响最大,符合理论研究结果。BP神经网络模型建立的成功为后续生活污水智能化控制的研究提供了相应的理论基础。     

Optimal reduction of the ozone monitoring network over France

Ozone is a harmful air pollutant at ground level, and its concentrations are measured with routine monitoring networks. Due to the heterogeneous nature of ozone fields, the spatial distribution of the ozone concentration measurements is very important. Therefore, the evaluation of distributed monitoring networks is of both theoretical and practical interests. In this study, we assess the efficiency of the ozone monitoring network over France (BDQA) by investigating a network reduction problem. We examine how well a subset of the BDQA network can represent the full network. The performance of a subnetwork is taken to be the root mean square error (rmse) of the hourly ozone mean concentration estimations over the whole network given the observations from that subnetwork. Spatial interpolations are conducted for the ozone estimation taking into account the spatial correlations. Several interpolation methods, namely ordinary kriging, simple kriging, kriging about the means, and consistent kriging about the means, are compared for a reliable estimation. Exponential models are employed for the spatial correlations. It is found that the statistical information about the means improves significantly the kriging results, and that it is necessary to consider the correlation model to be hourly-varying and daily stationary. The network reduction problem is solved using a simulated annealing algorithm. Significant improvements can be obtained through these optimizations. For instance, removing optimally half the stations leads to an estimation error of the order of the standard observational error (10 μg m^?3). The resulting optimal subnetworks are dense in urban agglomerations around Paris (Île-de-France) and Nice (Côte d’Azur), where high ozone concentrations and strong heterogeneity are observed. The optimal subnetworks are probably dense near frontiers because beyond these frontiers there is no observation to reduce the uncertainty of the ozone field. For large rural regions, the stations are uniformly distributed. The fractions between urban, suburban and rural stations are rather constant for optimal subnetworks of larger size (beyond 100 stations). By contrast, for smaller subnetworks, the urban stations dominate.     

Simulations of the dispersion of reactive pollutants in a street canyon,considering different chemical mechanisms and micromixing

The Stochastic Fields (SF) or Field Monte Carlo method has been used to model the dispersion of reactive scalars in a street canyon, using a simple chemistry and the CBM-IV mechanism. SF is a Probability Density Function (PDF) method which allows both means and variances of the scalars to be calculated as well as considering the effect of segregation on reaction rates. It was found that the variance of reactive scalars such as NO₂ was very high in the mixing region at roof-top level with rms values of the order of the mean values. The effect of segregation on major species such as O₃ was found to be very small using either mechanism, however, some radical species in CBM-IV showed a significant difference. These were found to be the seven species with the fastest chemical timescales. The calculated photostationary state defect was also found to be in error when segregation is neglected.     

Simulating industrial emissions using Atmospheric Dispersion Modeling System: Model performance and source emission factors

In this paper, the Gaussian Atmospheric Dispersion Modeling System (ADMS4) was coupled with field observations of surface meteorology and concentrations of several air quality indicators (nitrogen oxides (NO_X), carbon monoxide (CO), fine particulate matter (PM₁₀) and sulfur dioxide (SO₂)) to test the applicability of source emission factors set by the European Environment Agency (EEA) and the United States Environmental Protection Agency (USEPA) at an industrial complex. Best emission factors and data groupings based on receptor location, type of terrain and wind speed, were relied upon to examine model performance using statistical analyses of simulated and observed data. The model performance was deemed satisfactory for several scenarios when receptors were located at downwind sites with index of agreement d values reaching 0.58, fractional bias “FB” and geometric mean bias “MG” values approaching 0 and 1, respectively, and normalized mean square error “NMSE” values as low as 2.17. However, median ratios of predicted to observed concentrations “Cp/Co” at variable downstream distances were 0.01, 0.36, 0.76 and 0.19 for NO_X, CO, PM₁₀ and SO₂, respectively, and the fraction of predictions within a factor of two of observations “FAC2” values were lower than 0.5, indicating that the model could not adequately replicate all observed variations in emittant concentrations. Also, the model was found to be significantly sensitive to the input emission factor bringing into light the deficiency in regulatory compliance modeling which often uses internationally reported emission factors without testing their applicability.

Implications In the absence of site-specific source emission factors, the use of internationally reported emission factors without testing their validity may generate significant errors. Instead, recorded field measurements and meteorological data may be combined with atmospheric transport and dispersion models to better estimate source emissions, particularly in regulatory compliance studies. In this context, lower model performance is expected at higher wind speeds for most indicators such as CO, PM₁₀, and SO₂.     

Selection of an appropriate model to predict plume dispersion in coastal areas

In order to suggest a new methodology for selecting an appropriate dispersion model, various statistical measures having respective characteristics and recommended value ranges were integrated to produce a new single index by using fuzzy inference where eight statistical measures for various model results, including fractional bias (FB), normalized mean square error (NMSE), geometric bias mean (MG), geometric bias variance (VG), within a factor of two (FAC2), index of agreement (IOA), unpaired accuracy of the peak concentration (UAPC), and mean relative error (MRE), were taken as premise part variables. The new methodology using a single index was applied to the prediction of ground-level SO₂ concentration of 1-h average in coastal areas, where eight modeling combinations were organized with fumigation models, σ_y schemes for pre-fumigation, and modification schemes for σ_y during fumigation. As a result, the fumigation model of Lyons and Cole was found to have better predictability than the modified Gaussian model assuming that whole plume is immerged into the Thermal Internal Boundary Layer (TIBL). Again, a better scheme of σ_y (fumigation) was discerned. This approach, which employed the new integrated index, appears to be applicable to model evaluation or selection in various areas including complex coastal areas.