基于小波变换与传统时间序列模型的臭氧浓度多步预测

采用最大重叠小波分解与重构方法,将影响O3小时浓度的不同时间尺度的物化过程分离出来,以提高序列的光滑性.同时,选择合适的传统时间序列模型(如ARIMA模型等)来描述不同过程的序列特征,并分别拟合预报.最后,在建模中引入24 h季节项,以实现提前24 h-次性预测未来1d的O3逐时浓度.结果表明,预报的平均相对误差为12.92％,平均绝对误差和均方根误差分别为10.04 μ.g·m-3和13.98μg·m-3,预报值与实测值的相关系数和匹配指数分别为0.96和0.98.随着预测期的延长,预报误差仍处于可接受范围内.该方法同样适用于每日最大O3小时浓度预报,研究结果为发布天气预报式的空气质量预报提供了新思路,便于公众规划出行并减少大气污染暴露.     

COMPARISON OF REGRESSION AND TIME-SERIES METHODS FOR SYNTHESIZING MISSING STREAMFLOW RECORDS1

ABSTRACT: Regression and time-series techniques have been used to synthesize and predict the stream flow at the Foresta Bridge gage from information at the upstream Pohono Bridge gage on the Merced River near Yosemite National Park. Using the available data from two time periods (calendar year 1979 and water year 1986), we evaluated the two techniques in their ability to model the variation in the observed flows and in their ability to predict stream flow at the Foresta Bridge gage for the 1979 time period with data from the 1986 time period. Both techniques produced reasonably good estimates and forecasts of the flow at the downstream gage. However, the regression model was found to have a significant amount of autocorrelation in the residuals, which the time-series model was able to eliminate. The time-series technique presented can be of great assistance in arriving at reasonable estimates of flow in data sets that have large missing portions of data.     

基于ARIMA的安全生产事故预测与实证研究

为了减少企业安全管理者在生产作业中由于不确定性导致错误决策所产生的风险,在企业历年安全生产事故数据基础上进行预测具有一定的现实意义。以某企业2008年至2011年的安全生产事故次数时序数据,采用EViews 5统计分析软件,基于ARIMA时间序列预测模型更加关注对事故发生是否平稳而相对于其他预测模型更关注于趋势研究的良好特点,建立安全生产事故ARIMA时序预测模型,并对2012年的安全生产事故发生次数进行预测,通过效果检验发现该模型预测结果基本上能够反映该企业安全生产事故发生的实际情况。通过ARIMA方法在某企业安全生产事故预测具体案例的实现,是对现有安全生产事故预测方法的补充和完善,可为企业安全管理和决策提供一定的指导。     

基于ARIMA和BP神经网络组合模型的我国碳排放强度预测

预测我国碳排放强度的长期变动趋势, 对国家进行宏观经济管理和节能减排工作具有重要的参考价值。运用深入分析自回归移动平均模型和神经网络的特性，并在此基础上建立ARIMA模型和BP神经网络组合模型，将碳排放强度的时间序列的数据结构分解为线性和非线性残差部分，对我国碳排放强度的变化趋势进行了综合分析与预测。结果显示:今后10 a我国碳排放强度总体是逐步下降的，但到2020年我国碳排放强度仅比2005年下降34%，比我国政府提出碳排放强度下降40%~45%的目标还有一定的差距。因此，要在2020年实现我国碳排放强度目标，必须要调整宏观经济政策，采取各种政策措施以实现目标     

The Iowa Graduated Driver Licensing program: effectiveness in reducing crashes of teenage drivers

PROBLEM: Graduated Driver Licensing (GDL) programs vary in the United States in terms of implementation and restrictions. The State of Iowa's GDL program is assessed for its effectiveness in reducing crashes among teenage drivers. METHOD: Time series analysis was used to evaluate police documented crashes involving 16-, 17-, and 18-year-old drivers over a 10 year period, with an intervention identified at the point of GDL implementation. RESULTS: After controlling for seasonal trends and auto-correlative effects, a significant reduction in the crash rate of and 16- and 17-year-old drivers was observed due to the GDL implementation. However, there were no significant reductions in crash rates for 18-year-old drivers. DISCUSSION: The analyses suggest that the Iowa GDL program is effective in reducing the crash rates of 16- and 17-year-old drivers but the effects do not sustain for 18-year-old drivers. IMPACT ON INDUSTRY: The results suggest that the program appears to be working, however further analysis is needed to determine what factors are preventing lasting effects for these teenage drivers.     

Effectiveness of lowering the blood alcohol concentration (BAC) limit for driving from 0.10 to 0.08 grams per deciliter in the United States

Objective: The current study evaluates of the effects of lowering the blood alcohol concentration (BAC) limit from 0.10 to 0.08?g/dL across all 50 states in the United States. Our objectives were to (1) estimate the effects of the 0.08?g/dL BAC limit on drinking driver fatal crash rates; (2) compare the effects from early-adopting states to the effects of late-adopting states; (3) determine the effects on drivers with low BACs (0.01–0.07?g/dL) and high BACs (0.08+ g/dL); and (4) estimate the lives saved since 1983 due to the adoption of 0.08?g/dL BAC laws.

Methods: Our study examined annual data from the Fatality Analysis Reporting System (FARS) for each jurisdiction from 1982 through 2014. Our basic outcome measure was the ratio of drinking drivers (BAC ≥0.01?g/dL) to nondrinking drivers (BAC?=?0.00?g/dL). Covariates included 0.10 BAC laws, administrative license revocation (ALR) laws, seat belt laws, minimum legal drinking age (MLDA) laws, and unemployment rates. We utilized autoregressive integrated moving average (ARIMA) models for each state, where the implementation date of the law was modeled as a zero-order transfer function in the series, in addition to any extant trends that may have been occurring simultaneously. Before determining the specific impact of the implementation of 0.08?g/dL BAC laws, we conducted a time series analysis for each state. We tested for between-state mediating factors relating to our covariates.

Results: A total of 38 of the 51 jurisdictions showed that lowering the BAC limit was associated with reduced drinking driver fatal crash ratios, with 20 of those reductions being significant. The total effects showed a 10.4% reduction in annual drinking driver fatal crash rates, which is estimated to have saved an average of 1,736 lives each year between 1983 and 2014 and 24,868 lives in total. Implementing a BAC limit of 0.08?g/dL had significant impacts on both high- and low-BAC fatal crash ratios. Though early-adopting jurisdictions (1983–1999) demonstrated a larger decrease in fatal drinking driver crash ratios than did late-adopting jurisdictions (2000–2005), the results were not statistically significant (P?>?.05).

Conclusions: Our study of the effects of lowering the BAC from 0.10 to 0.08?g/dL in the United States from 1982 to 2014 showed an overall effect of 10.4% on annual drinking driver fatal crash rates, in line with other multistate studies. This research provides strong evidence of the relationship between lowering the BAC limit for driving and the general deterrent effect on impaired-driving fatal crash rates.     

建筑工程生产事故死亡人数时间序列分析

为研究建筑工程安全生产事故死亡人数的变化规律,采用时间序列分析方法,分析了建筑安全事故死亡人数时间序列上的趋势性规律,通过数据预处理和模型的识别与检验,最终建立了安全事故死亡人数预测模型。对全国2005—2014年建筑工程安全生产事故造成的死亡人数进行了分析和预测。结果表明:ARIMA模型各年预测值与实际值误差率为0.393,相比灰色模型和BP神经网络模型误差率最小。总体上说,ARIMA模型较适用于随机性较大的数据的趋势预测。     

基于ARIMA-LOESS模型的危化品道路运输事故起数预测

文章针对危化品道路运输事故预测问题,运用差分自回归移动平均模型(Autoregressive Integrated Moving Average,ARIMA)与局部加权回归模型(Locally Estimated Scatterplot Smoothing,LOESS)的组合模型,对我国危化品道路运输事故发生起数进行预测。首先,基于2011—2018年我国发生的危化品道路运输事故数据建立ARIMA模型,利用SPSS软件进行模型拟合预测,获取危化品道路运输事故起数的线性部分;其次,应用MATLAB建立LOESS回归模型,对ARIMA模型预测偏差进行残差优化,获取危化品道路运输事故起数的非线性部分;最后,建立ARIMA-LOESS组合模型,利用组合模型对危化品道路运输事故发生起数进行预测,并根据真实数据对预测结果进行对比验证。结果表明:ARIMA-LOESS组合预测模型可较好拟合危化品道路运输事故数据序列,并修正单一模型的误差,获取较高的预测精度。该研究可为危化品道路运输安全与运行的趋势分析与判断提供更加可靠的数据依据,也可为危化品道路运输事故防控方案提供帮助。     

MULTIPLICATIVE,SEASONAL ARIMA MODELS FOR LAKE ERIE AND LAKE ONTARIO WATER LEVELS1

ABSTRACT: Federal agencies in the U.S. and Canada continuously examine methods to improve understanding and forecasting of Great Lakes water level dynamics in an effort to reduce the negative impacts of fluctuating levels incurred by interests using the lakes. The short term, seasonal and long term water level dynamics of lakes Erie and Ontario are discussed. Multiplicative, seasonal ARIMA models are developed for lakes Erie and Ontario using standardized, monthly mean level data for the period 1900 to 1986. The most appropriate model identified for each lake had the general form: (1 0 1)(0 1 1)₁₂. The data for each lake were subdivided by time periods (1900 to 1942;1 943 to 1986) and the model coefficients estimated for the subdivided data were similar, indicating general model stability for the entire period of record. The models estimated for the full data sets were used to forecast levels 1,2,3, and 6 months ahead for a period of high levels (1984 to 1986). The average absolute forecast error for Lake Erie was 0.049m, 0.076m, 0.091 m and 0.128m for the 1, 2,3, and 6 month forecasts, respectively. The average absolute forecast error for Lake Ontario was 0.058m, 0.095m, 0.120m and 0.136m for the 1,2,3, and 6 month forecasts, respectively. The ARIMA models provide additional information on water level time series structure and dynamics. The models also could be coordinated with current forecasting methods, possibly improving forecasting accuracy.     

基于ARIMA与LS-SVM组合模型的飞行事故征候预测

应用差分自回归移动平均模型(ARIMA)和最小二乘支持向量机(LS-SVM)的组合模型,对某航空公司的月度事故征候万时率进行了预测分析。对2008—2016年某航空公司的事故征候、飞行小时、航空器数量等历史数据建立ARIMA模型,应用SPSS软件进行模型拟合,获得事故征候万时率的线性部分;随后利用LS-SVM分析ARIMA模型的残差,获取非线性部分,最终通过二者之和获得ARIMA+LSSVM组合模型。对2017年1—3月的月度事故征候万时率进行了预测,并用实际数据验证。结果表明:ARIMA(1,1,1)(1,1,1)12模型较好地拟合了事故征候万时率的历史序列,LS-SVM模型对残差的拟合获得了较好的精度;组合模型的短期(3个月)预测值与航空公司事故征候万时率的趋势完全一致,且预测精确度可接受。