Exchange rate forecasting with optimum singular spectrum analysis

Forecasting exchange rate is undoubtedly an attractive and challenging issue that has been of interest in different domains for many years. The singular spectrum analysis （SSA） technique has been used as a promising technique for time series forecasting including exchange rate series. The SSA technique is based upon two main choices： Window length, L, and the number of singular values, r. These values are very important for the reconstruction stage and forecasting purposes. Here the authors consider an optimum version of the SSA technique for forecasting exchange rates. The forecasting performances of the SSA technique for one-step-ahead forecast of six exchange rate series are used to find the best L and r.     

Estimating multi-country prosperity index: A two-dimensional singular spectrum analysis approach

With the development of the global economy, interaction among different economic entities from one region is intensifying, which makes it significant to consider such interaction when constructing composite index for each country from one region. Recent advances in signal extraction and time series analysis have made such consideration feasible and practical. Singular spectrum analysis （SSA） is a well-developed technique for time series analysis and proven to be a powerful tool for signal extraction. The present study aims to introduce the usage of the SSA technique for multi-country business cycle analysis. The multivariate SSA （MSSA） is employed to construct a model-based composite index and the two dimensional SSA （2D-SSA） is employed to establish the multi-country composite index. Empirical results performed on Chinese economy demonstrate the accuracy and stability of MSSA-based composite index, and the 2D-SSA based composite indices for Asian countries confirm the efficiency of the technique in capturing the interaction among different countries.     

A DECOMPOSITION METHOD OF STRUCTURAL DECOMPOSITION ANALYSIS

Over the past two decades, structural decomposition analysis (SDA) has developed into a major analytical tool in the field of input-output (IO) techniques, but the method was found to suffer from one or more of the following problems. The decomposition forms, which are used to measure the contribution of a specific determinant, are not unique due to the existence of a multitude of equivalent forms, irrational due to the weights of different determinants not matching, inexact due to the existence of large interaction terms.In this paper, a decomposition method is derived to overcome these deficiencies, and we prove that the result of this approach is equal to the Shapley value in cooperative games,and so some properties of the method are obtained. Beyond that, the two approaches that have been used predominantly in the literature have been proved to be the approximate solutions of the method.     

A transfer forecasting model for container throughput guided by discrete PSO

Abstract Accurate forecast of future container throughput of a port is very important for its con struction, upgrading, and operation management. This study proposes a transfer forecasting model guided by discrete particle swarm optimization algorithm （TF-DPSO）. It firstly transfers some related time series in source domain to assist in modeling the target time series by transfer learning technique, and then constructs the forecasting model by a pattern matching method called analog complexing. Finally, the discrete particle swarm optimization algorithm is introduced to find the optimal match between the two important parameters in TF-DPSO. The container throughput time series of two im portant ports in China, Shanghai Port and Ningbo Port are used for empirical analysis, and the results show the effectiveness of the proposed model.     

Series expansion feasibility of singular integral in method of moments

When calculating electromagnetic scattering using method of moments （MoM）, integral of the singular term has a significant influence on the results. This paper transforms the singular surface integral to the contour integral. The integrand is expanded to Taylor series and the integral results in a closed form. The cut-off error is analyzed to show that the series converges fast and only about 2 terms can agree wel with the accurate result. The comparison of the perfect electric conductive （PEC） sphere＇s bi-static radar cross section （RCS） using MoM and the accurate method validates the feasibility in manipulating the singularity. The error due to the facet size and the cut-off terms of the series are analyzed in examples.     

Exploring Long-Memory Process in the Prediction of Interval-Valued Financial Time Series and Its Application

Long-memory process has been widely studied in classical financial time series analysis,which has merely been reported in the field of interval-valued financial time series. The aim of this paper is to explore long-memory process in the prediction of interval-valued time series(Iv TS). To model the long-memory process, two novel interval-valued time series prediction models named as intervalvalued vector autoregressive fractionally integrated moving average(IV-VARFIMA) and ARFIMAXFIGARCH were es...     

Learning multivariate time series causal graphs based on conditional mutual information

Detection and clarification of cause-effect relationships among variables is an important problem in time series analysis.This paper provides a method that employs both mutual information and conditional mutual information to identify the causal structure of multivariate time series causal graphical models.A three-step procedure is developed to learn the contemporaneous and the lagged causal relationships of time series causal graphs.Contrary to conventional constraint-based algorithm, the proposed algorithm does not involve any special kinds of distribution and is nonparametric.These properties are especially appealing for inference of time series causal graphs when the prior knowledge about the data model is not available.Simulations and case analysis demonstrate the effectiveness of the method.     

Measuring Nonlinear Dependence Between Time Series Based on Correlation Dimension

In this paper, it is proved that the correlation dimension estimate of a nonlinear dynamical system with its multivariate observation series is the same as that with its univariate observation series. Based on this result, an inference method is presented, and the Nonlinear Dependence Coefficient is defined. This method is designed for testing nonlinear dependence between time series, and can be used in economic analysis and forecasting. Numerical results show the method is effective.     

AN IMPROVED MULTIVARIATE LOSS FUNCTION APPROACH TO OPTIMIZATION

The basic purpose of a quality loss function is to evaluate a loss to customers in a quantitativemanner.Although there are several multivariate loss functions that have been proposed and studied inthe literature,it has room for improvement.A good multivariate loss function should represent anappropriate compromise in terms of both process economics and the correlation structure amongvarious responses.More important,it should be easily understood and implemented in practice.According to this criterion,we first introduce a pragmatic dimensionless multivariate loss functionproposed by Artiles-Leon,then we improve the multivariate loss function in two respects:one ismaking it suitable for all three types of quality characteristics;the other is considering correlationstructure among the various responses,which makes the improved multivariate loss function moreadequate in the real world.On the bases of these,an example from industrial practice is provided tocompare our improved method with other methods,and la     

基于奇异谱分析的我国航空客运量集成预测模型

针对时间序列包含噪声以及单一模型可能存在预测表现不稳定的问题,本文提出了一个基于奇异谱分析(SSA)的集成预测模型,并将其运用于我国年度航空客运量的预测中.首先,采用SSA方法对原始时间序列进行分解和重构,得到一个剔除噪声的时间序列,然后将其作为单整自回归移动平均模型(ARIMA)、支持向量回归模型(SVR)、Holt-Winters方法(HW)等单一模型的输入并进行预测,接着再采用加权平均集成预测方法(WA)将三种单一模型的预测结果进行综合集成.通过与各单一模型、基于经验模态分解方法(EMD)的模型以及简单平均集成预测方法(SA)的预测结果进行对比发现,本文所建模型具有较高的预测精度和较稳定的预测表现.最后,采用本文的模型对我国2014-2016年年度航空客运量进行了预测.     

多变量混沌时间序列Volterra自适应实时预测

针对多变量混沌时间序列,给出一种Volterra滤波器实现结构.该滤波器利用基于奇异值分解的最小二乘法确定初始核,通过归一化最小均方差(normalized least mean square,NLMS)算法实时确定滤波系数,并用这种多变量Volterra结构对Lorenz时间序列进行仿真.计算结果表明,在无噪声情况下,该方法的实时一步预测精度比目前单变量混沌时间序列Volterra自适应预测方法的一步预测精度提高了102倍,表明这种实现结构易实现且收敛性能更好;在有噪声的情况下,该方法的实时多步预测性能优于局部多项式预测法的多步预测性能,且抗噪性更强.     

Singular spectrum analysis: methodology and application to economics data

This paper describes the methodology of singular spectrum analysis (SSA) and demonstrate that it is a powerful method of time series analysis and forecasting, particulary for economic time series. The authors consider the application of SSA to the analysis and forecasting of the Iranian national accounts data as provided by the Central Bank of the Islamic Republic of Iran. This research was in part supported by a grant (No. 88/121230) from Institute for Trade Studies and Research (ITSR), Tehran, Iran.     

基于奇异谱分析的航空客运需求分析与分解集成预测模型

考虑到航空旅客运输需求影响因素复杂以及航空客运需求序列非线性非平稳等特征,本文提出了一个基于奇异谱分析(SSA)的航空客运需求分析与分解集成预测模型.需求分析阶段,首先使用SSA对航空客运需求序列进行有效分解,接着借助奇异熵理论,将序列重构为长期趋势项、中期市场波动项和短期噪声项;预测阶段,使用排列熵(PE)判断各重构序列复杂度的高低,并依据序列复杂度分别选择粒子群算法(PSO)和布谷鸟算法(CS)双优化的支持向量回归模型(SVR)或单整自回归移动平均模型(ARIMA)进行预测,结果表明,该分解集成预测模型较ARIMA、SVR等基准模型有着更好的预测性能.     

基于EOF-SVD模型的多元时间序列相关性研究及预测

将奇异值分解同自然正交分解相结合,提出一种改进的正交奇异值分解方法.通过对原始数据进行自然正交分解,削弱原始数据之间的相关性,增强其用于分析及预测的能力,并得到相互正交的主成分代替原始数据进行奇异值分解,分析两个变量场之间的相关关系.在此基础上建立神经网络预测模型,实现多元时间序列的预测.采用该方法对三门峡处径流量同太平洋海温的耦合关系进行分析,并同常规奇异值分解方法进行比较,仿真结果验证了所提方法的有效性.     

事件影响下的时间序列多尺度集成预测

针对时间序列的非线性、非平稳和多尺度特征,考虑到事件对序列结构产生的影响,提出事件影响下的时间序列多尺度集成预测策略。首先,基于经验模态分解将原始序列分解成若干分量,从多个尺度展现序列的基本构成;随后,基于迭代累积平方和实现分量序列的变点检验,从多个尺度判别和获取事件对序列产生的结构性影响;然后,基于干预分析构建事件对不同分量序列的影响模型,据此剔除事件影响,获取净化序列;最后,运用基于粒子群优化的支持向量回归,建立单一尺度的序列预测模型,进而实现事件影响下的时间序列多尺度集成预测。实证分析表明:该策略能够精细辨识事件对序列的多尺度影响,有效建立序列总体及分量的预测模型,与传统方法相比,具有更强的事件辨识能力、自适应建模能力和更高的预测精度。     

Measuring and locating zones of chaos and irregularity

The new measures computed here are the spectral detrended fluctuation anatysls （sDFA） and spectral multi-taper method （sMTM）. sDFA applies the standard detrended fluctuation analysis （DFA） algorithm to power spectra, sMTM exploits the minute increases in the broadband response, typical of chaotic spectra approaching optimal values. The authors chose the Brusselator, Lorenz, and Duffing as the proposed models to measure and locate chaos and severe irregularity. Their series of chaotic parametric responses in short time-series is advantageous. Where cycles have only a limited number of slow oscillations such as for systems biology and medicine. It is difficult to create, locate, or monitor chaos. From 50 linearly increasing starting points applied to the chaos target function （CTF）; the mean percentage increases in Kolmogorov-Sinai entropy （KS-Entropy） for the proposed chosen models; and p-values when the models were compared statistically by Kruskal-Wallis and ANOVA1 test with distributions assumed normal are Duffing （CTF： 31%： p 〈0.03）; Lorenz （CTF： 2%： p 〈0.03）, and I3russelator （CTF： 8%： p 〈0.01）. Principal component analysis （PCA） is applied to assess the significance of the objective functions for tuning the chaotic response. From PCA the conclusion is that CTF is the most beneficial objective function overall delivering the highest increases in mean KS-Entropy.     

基于相空间同步的多变量序列相关性分析及预测

针对多变量混沌序列相关性分析中各分量幅值之间可能没有明显的相关性,但在其相空间邻域内会产生同步特性的问题,提出一种从相空间同步角度研究两个变量间相互依赖关系的非线性相关分析方法。首先按照对应的时间标记将原始变量相空间中的邻域点向另外一个变量中进行投影,分析映射前后邻域半径的变化,在此基础上定义一种度量变量间非线性相关性的评价指标。最后构建多变量局域预测模型,实现对多变量混沌序列的精确预测。仿真实例验证了结果的有效性。