基于ARIMA和支撑矢量机组合预测的惯性器件故障预报

文中提出一种基于ARIMA和支撑矢量机进行组合预测的方法,该方法综合运用ARIMA良好的线性拟合能力和支撑矢量机强大的非线性映射功能,将两种预测模型有机地组合在一起,综合各自优点,以期有效改善模型的拟合能力,获得最优预测效果.文中将该方法应用于某飞行器惯性器件的故障预报当中并进行了仿真实验.结果表明,这种方法相对于单项模型的预测具有更高的精度,对于复杂时间序列的分析和预测有很好的应用价值,在故障预报中具有广泛的应用前景.     

Prediction of spectral reflectance factor distribution of automotive paint finishes

It is necessary to determine the accurate reflectance of painted surfaces for the review of paint finishes by computer graphics (CG) before actual painting of the exterior color of automobiles, and for quality control during production and inspection processes. We have optimized a method for measuring reflectance by using a statistical technique. We have found that the reflectance of a painted surface is best measured at an incident angle of 60° and at five aspecular angles of 10°, 18°, 28°, 40°, and 90°. Our method makes it possible to accurately reproduce reflection characteristics of paint finishes containing special flake pigments, such as pearl mica. Also it was proved that our method can apply not only to solid and metallic coatings but to all painted surfaces. © 2005 Wiley Periodicals, Inc. Col Res Appl, 30, 275–282, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/col.20125     

A Nonparametric Prewhitened Covariance Estimator

This paper proposes a new nonparametric spectral density estimator for time series models with general autocorrelation. The conventional nonparametric estimator that uses a positive kernel has mean squared error no better than n^?4/5. We show that the best implementation of our estimator has mean squared error of order n^?8/9, provided there is sufficient smoothness present in the spectral density. This is, of course, achieved by bias reduction; however, unlike most other bias reduction methods, like the kernel method with higher‐order kernels, our procedure ensures a positive definite estimate. Our method is a generalization of the well‐known prewhitening method of spectral estimation; we argue that this can best be interpreted as multiplicative bias reduction. Higher‐order expansions for the proposed estimator are derived, providing an improved bandwidth choice that minimizes the mean squared error to the second order. A simulation study shows that the recommended prewhitened kernel estimator reduces bias and mean squared error in spectral density estimation.     

Time series regression-based pairs trading in the Korean equities market

Abstract

Pairs trading is an instance of statistical arbitrage that relies on heavy quantitative data analysis to profit by capitalising low-risk trading opportunities provided by anomalies of related assets. A key element in pairs trading is the rule by which open and close trading triggers are defined. This paper investigates the use of time series regression to define the rule which has previously been identified with fixed threshold-based approaches. Empirical results indicate that our approach may yield significantly increased excess returns compared to ones obtained by previous approaches on large capitalisation stocks in the Korean equities market.     

Quality evaluation of ink-jet paper with principal components analysis

Non-impact printing techniques have in many fields of application replaced traditional printing methods. Apart from factors such as the type and conditions of ink/paper transfer and ink properties, the quality of the image produced with an office ink-jet printer depends decisively on paper characteristics. In this study, various medium- and photo-quality ink-jet paper sorts were examined for their general, optical, surface and printing properties. Based on these results, the ranking of tested paper properties according to their significance for ink-jet paper quality on one hand and the suitability of individual paper sorts for this printing technique on the other was determined using multivariate tool principal components analysis.     

单变量ARIMA模型的铁水钛含量预报

以某大型钢铁企业四号高炉为背景,针对其钛含量采样周期不固定的特点,提出了利用三次样条函数对采样数据序列进行等间距化处理,再利用ARIMA模型进行建模及预测的综合方案.利用Visual C++及SPSS对方案进行了模拟实验.结果表明:该方案预测精度高,符合工艺要求,实现简单,适于在线运行.     

基于季节ARIMA模型的近海表层水温时间序列模拟与预测

基于1996—2007年逐月时间序列数据,采用季节ARIMA模型对连云港近海表层水温时间序列进行模拟,并依据残差不相关和简洁性原则确定模型的结构,建立最优预测模型ARIMA（1,0,1）（0,1,0）12。运用该模型对2008年逐月表层水温进行预测,预测值与实际值的变动趋势基本一致,且平均相对误差仅为3.5%。在此基础上对2009年连云港近海逐月表层水温进行预报,预报结果符合该海域表层水温的逐月变化趋势,表明模型用于近海表层水温预报是可行的。     

基于ARIMA和Prophet的水质预测集成学习模型

将时间序列模型ARIMA和Prophet作为基学习器,结合BP神经网络模型构建了水质预测集成学习模型。选取长江流域某断面2019—2020年的DO、CODMn、NH₃-N、TP和TN等5个水质指标的监测数据对该模型的有效性进行了检验,结果表明:5个水质指标集成学习模型预测结果的平均绝对百分比误差比时间序列模型的预测误差分别低35.0%、29.9%、4.1%、40.6%和17.1%,模型预测值和监测值的皮尔逊相关系数大于0.8。集成学习模型预测精度高于单一模型,可以更精确地进行水质预测。     

四川入境游人数的预报

根据2004年1月至2010年3月四川入境游数据,运用时间序列分析方法建立干预模型,进而以所建立的干预模型对四川入境游人数进行了预测,为四川省将来入境游事业的规划与发展提供理论依据和政策建议.     

计及温度影响的短期负荷预测时间序列模型

时间序列模型在国际和国内的短期电力负荷预测中得到了广泛应用.然而,这种方法的一个主要缺点是无法将影响负荷预测的主要因素之一即气象因素考虑进去.在此背景下,首先基于负荷和气温数据建立了负荷预测的回归模型,然后构造了回归模型残差累积式自回归一滑动平均模型并对回归模型进行修正.最后,用广东电力系统的实际负荷数据说明了所发展的...