DOCSLIB.ORG

LECTURES 2 - 3 : Stochastic Processes, Autocorrelation Function

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
LECTURES 2 - 3 : Stochastic Processes, Autocorrelation Function LECTURES 2 - 3 : Stochastic Processes, Autocorrelation function. Stationarity. Important points of Lecture 1: A time series fXtg is a series of observations taken sequentially over time: xt is an observation recorded at a specific time t. Characteristics of times series data: observations are dependent, become available at equally spaced time points and are time-ordered. This is a discrete time series. The purposes of time series analysis are to model and to predict or forecast future values of a series based on the history of that series. 2.2 Some descriptive techniques. (Based on [BD] x1.3 and x1.4) ......................................................................................... Take a step backwards: how do we describe a r.v. or a random vector? ² for a r.v. X: 2 d.f. FX (x) := P (X · x), mean ¹ = EX and variance σ = V ar(X). ² for a r.vector (X1;X2): joint d.f. FX1;X2 (x1; x2) := P (X1 · x1;X2 · x2), marginal d.f.FX1 (x1) := P (X1 · x1) ´ FX1;X2 (x1; 1) 2 2 mean vector (¹1; ¹2) = (EX1; EX2), variances σ1 = V ar(X1); σ2 = V ar(X2), and covariance Cov(X1;X2) = E(X1 ¡ ¹1)(X2 ¡ ¹2) ´ E(X1X2) ¡ ¹1¹2. Often we use correlation = normalized covariance: Cor(X1;X2) = Cov(X1;X2)=fσ1σ2g ......................................................................................... To describe a process X1;X2;::: we define (i) Def. Distribution function: (fi-di) d.f. Ft1:::tn (x1; : : : ; xn) = P (Xt1 · x1;:::;Xtn · xn); i.e. this is the joint d.f. for the vector (Xt1 ;:::;Xtn ). (ii) First- and Second-order moments. ² Mean: ¹X (t) = EXt 2 2 2 2 ² Variance: σX (t) = E(Xt ¡ ¹X (t)) ´ EXt ¡ ¹X (t) 1 ² Autocovariance function: γX (t; s) = Cov(Xt;Xs) = E[(Xt ¡ ¹X (t))(Xs ¡ ¹X (s))] ´ E(XtXs) ¡ ¹X (t)¹X (s) (Note: this is an infinite matrix). ² Autocorrelation function: Cov(Xt;Xs) γX (t; s) ½X (t; s) = Cor(Xt;Xs) = p = V ar(Xt)V ar(Xs) σX (t)σX (s) Properties of the process which are determined by the first- and second- order moments are called second-order properties. Example: stationarity. – will be defined soon! 2.3 Meaning of the Autocorrelation function: In linear regression problem, PSTAT 5A : Cor(X; Y ) measures linear dependence. Recall: correlation is zero if there is no linear dependence and it is close to §1 for variables with high linear dependence. In TS: Autocorrelation function is used to assess numerically the dependence between two adjacent values. Let say that for a moment we observe only two r. v.’s: Xt and Xt+1. 2 ½(t; t + 1) is the correlation of Xt;Xt+1. The prefix auto is to convey the notion of self- correlation (both variables come from the same TS), correlation of the series with itself. When the TS is smooth, the autocorr. f’n is large even when t and s are quite apart, whereas very choppy series tend to have autocorr. f’ns that are nearly zero for large separations. So, the autocorr. f’n tends to reflect the essential smoothness of the TS. Let us consider two “extreme” examples: Example: Gaussian White Noise: Consider the series fZtg, where Zt are Gaussian, m.z. r.v.’s and γZ (s; t) = E(ZtZs) = 1; if s = t ; and 0; if s 6= t so that the autocov. f’n is 0 at all time separations. The series looks very choppy: 1 Example. Apply smoothing operation Xt = 3 (Zt¡1 + Zt + Zt+1). Autocov. f’n is: Autocor. f’n is a normalized autocov. f’n. 3. Stationarity 3.1 Stationarity and Strict Stationarity (Based on x1.4 and 2.1 of [BD]) Intuitively stationarity means that the graphs over two equal-length time intervals of a realization of the TS should exibit similar statistical characteristics. 3 On a graph: ² no trend ² no seasonality ² no change of variability ² no apparent sharp changes of behavior Two approaches to stationarity: (i) Def. A stochastic process is said to be strictly stationary if the joint probability density associated with the n r.v.’s Xt1 ;:::;Xtn for any set of times t1; : : : ; tn, is the same as that associated with the n r.v.’s Xt1+k;:::;Xtn+k for any integer k. (ii) Def. A process is said to be (weakly, second-order) stationary if 2 (a) EjXtj < 1 (b) EXt = ¹ for all t 2 T (c) γX (t; s) = γX (t + r; s + r) = γX (t ¡ s) for all t; s; r 2 T . Note: The function γX (k) is often refered as the value of the autocov. f’n at lag k. (iii) The ACF (the autocorr. f’n) is γX (k) ½X (k) = γX (0) 2 (iv) Note: γX (k) = γX (¡k); ½X (k) = ½X (¡k); γX (0) = σX . (v) Example: Apply smoothing operation to WN series: 1 Xt = 3 (Zt¡1 + Zt + Zt+1). Autocov. f’n is: 2 (vi) Example: Xt = Z1 + ::: + Zt, where Zt are i.i.d. r.v.s with m.z. and variance σ . Non-stationary. 4 In general: ² strict stationarity + finite second moment imply stationarity In general: ² stationarity does not imply strict stationarity Fact: For Gaussian TS stationarity = strict stationarity. (because Gaussian distribution is determined by its mean and covariance) 5.
Load more

Recommended publications
  • Power Comparisons of Five Most Commonly Used Autocorrelation Tests
    Pak.j.stat.oper.res. Vol.16 No. 1 2020 pp119-130 DOI: http://dx.doi.org/10.18187/pjsor.v16i1.2691 Pakistan Journal of Statistics and Operation Research Power Comparisons of Five Most Commonly Used Autocorrelation Tests Stanislaus S. Uyanto School of Economics and Business, Atma Jaya Catholic University of Indonesia, [email protected] Abstract In regression analysis, autocorrelation of the error terms violates the ordinary least squares assumption that the error terms are uncorrelated. The consequence is that the estimates of coefficients and their standard errors will be wrong if the autocorrelation is ignored. There are many tests for autocorrelation, we want to know which test is more powerful. We use Monte Carlo methods to compare the power of five most commonly used tests for autocorrelation, namely Durbin-Watson, Breusch-Godfrey, Box–Pierce, Ljung Box, and Runs tests in two different linear regression models. The results indicate the Durbin-Watson test performs better in the regression model without lagged dependent variable, although the advantage over the other tests reduce with increasing autocorrelation and sample sizes. For the model with lagged dependent variable, the Breusch-Godfrey test is generally superior to the other tests. Key Words: Correlated error terms; Ordinary least squares assumption; Residuals; Regression diagnostic; Lagged dependent variable. Mathematical Subject Classification: 62J05; 62J20. 1. Introduction An important assumption of the classical linear regression model states that there is no autocorrelation in the error terms. Autocorrelation of the error terms violates the ordinary least squares assumption that the error terms are uncorrelated, meaning that the Gauss Markov theorem (Plackett, 1949, 1950; Greene, 2018) does not apply, and that ordinary least squares estimators are no longer the best linear unbiased estimators.
    [Show full text]
  • Autocovariance Function Estimation Via Penalized Regression
    Autocovariance Function Estimation via Penalized Regression Lina Liao, Cheolwoo Park Department of Statistics, University of Georgia, Athens, GA 30602, USA Jan Hannig Department of Statistics and Operations Research, University of North Carolina, Chapel Hill, NC 27599, USA Kee-Hoon Kang Department of Statistics, Hankuk University of Foreign Studies, Yongin, 449-791, Korea Abstract The work revisits the autocovariance function estimation, a fundamental problem in statistical inference for time series. We convert the function estimation problem into constrained penalized regression with a generalized penalty that provides us with flexible and accurate estimation, and study the asymptotic properties of the proposed estimator. In case of a nonzero mean time series, we apply a penalized regression technique to a differenced time series, which does not require a separate detrending procedure. In penalized regression, selection of tuning parameters is critical and we propose four different data-driven criteria to determine them. A simulation study shows effectiveness of the tuning parameter selection and that the proposed approach is superior to three existing methods. We also briefly discuss the extension of the proposed approach to interval-valued time series. Some key words: Autocovariance function; Differenced time series; Regularization; Time series; Tuning parameter selection. 1 1 Introduction Let fYt; t 2 T g be a stochastic process such that V ar(Yt) < 1, for all t 2 T . The autoco- variance function of fYtg is given as γ(s; t) = Cov(Ys;Yt) for all s; t 2 T . In this work, we consider a regularly sampled time series f(i; Yi); i = 1; ··· ; ng. Its model can be written as Yi = g(i) + i; i = 1; : : : ; n; (1.1) where g is a smooth deterministic function and the error is assumed to be a weakly stationary 2 process with E(i) = 0, V ar(i) = σ and Cov(i; j) = γ(ji − jj) for all i; j = 1; : : : ; n: The estimation of the autocovariance function γ (or autocorrelation function) is crucial to determine the degree of serial correlation in a time series.
    [Show full text]
  • Alternative Tests for Time Series Dependence Based on Autocorrelation Coefficients
    Alternative Tests for Time Series Dependence Based on Autocorrelation Coefficients Richard M. Levich and Rosario C. Rizzo * Current Draft: December 1998 Abstract: When autocorrelation is small, existing statistical techniques may not be powerful enough to reject the hypothesis that a series is free of autocorrelation. We propose two new and simple statistical tests (RHO and PHI) based on the unweighted sum of autocorrelation and partial autocorrelation coefficients. We analyze a set of simulated data to show the higher power of RHO and PHI in comparison to conventional tests for autocorrelation, especially in the presence of small but persistent autocorrelation. We show an application of our tests to data on currency futures to demonstrate their practical use. Finally, we indicate how our methodology could be used for a new class of time series models (the Generalized Autoregressive, or GAR models) that take into account the presence of small but persistent autocorrelation. _______________________________________________________________ An earlier version of this paper was presented at the Symposium on Global Integration and Competition, sponsored by the Center for Japan-U.S. Business and Economic Studies, Stern School of Business, New York University, March 27-28, 1997. We thank Paul Samuelson and the other participants at that conference for useful comments. * Stern School of Business, New York University and Research Department, Bank of Italy, respectively. 1 1. Introduction Economic time series are often characterized by positive
    [Show full text]
  • 3 Autocorrelation
    3 Autocorrelation Autocorrelation refers to the correlation of a time series with its own past and future values. Autocorrelation is also sometimes called “lagged correlation” or “serial correlation”, which refers to the correlation between members of a series of numbers arranged in time. Positive autocorrelation might be considered a specific form of “persistence”, a tendency for a system to remain in the same state from one observation to the next. For example, the likelihood of tomorrow being rainy is greater if today is rainy than if today is dry. Geophysical time series are frequently autocorrelated because of inertia or carryover processes in the physical system. For example, the slowly evolving and moving low pressure systems in the atmosphere might impart persistence to daily rainfall. Or the slow drainage of groundwater reserves might impart correlation to successive annual flows of a river. Or stored photosynthates might impart correlation to successive annual values of tree-ring indices. Autocorrelation complicates the application of statistical tests by reducing the effective sample size. Autocorrelation can also complicate the identification of significant covariance or correlation between time series (e.g., precipitation with a tree-ring series). Autocorrelation implies that a time series is predictable, probabilistically, as future values are correlated with current and past values. Three tools for assessing the autocorrelation of a time series are (1) the time series plot, (2) the lagged scatterplot, and (3) the autocorrelation function. 3.1 Time series plot Positively autocorrelated series are sometimes referred to as persistent because positive departures from the mean tend to be followed by positive depatures from the mean, and negative departures from the mean tend to be followed by negative departures (Figure 3.1).
    [Show full text]
  • Spatial Autocorrelation: Covariance and Semivariance Semivariance
    Spatial Autocorrelation: Covariance and Semivariancence Lily Housese ­P eters GEOG 593 November 10, 2009 Quantitative Terrain Descriptorsrs Covariance and Semivariogram areare numericc methods used to describe the character of the terrain (ex. Terrain roughness, irregularity) Terrain character has important implications for: 1. the type of sampling strategy chosen 2. estimating DTM accuracy (after sampling and reconstruction) Spatial Autocorrelationon The First Law of Geography ““ Everything is related to everything else, but near things are moo re related than distant things.” (Waldo Tobler) The value of a variable at one point in space is related to the value of that same variable in a nearby location Ex. Moranan ’s I, Gearyary ’s C, LISA Positive Spatial Autocorrelation (Neighbors are similar) Negative Spatial Autocorrelation (Neighbors are dissimilar) R(d) = correlation coefficient of all the points with horizontal interval (d) Covariance The degree of similarity between pairs of surface points The value of similarity is an indicator of the complexity of the terrain surface Smaller the similarity = more complex the terrain surface V = Variance calculated from all N points Cov (d) = Covariance of all points with horizontal interval d Z i = Height of point i M = average height of all points Z i+d = elevation of the point with an interval of d from i Semivariancee Expresses the degree of relationship between points on a surface Equal to half the variance of the differences between all possible points spaced a constant distance apart
    [Show full text]
  • Modeling and Generating Multivariate Time Series with Arbitrary Marginals Using a Vector Autoregressive Technique
    Modeling and Generating Multivariate Time Series with Arbitrary Marginals Using a Vector Autoregressive Technique Bahar Deler Barry L. Nelson Dept. of IE & MS, Northwestern University September 2000 Abstract We present a model for representing stationary multivariate time series with arbitrary mar- ginal distributions and autocorrelation structures and describe how to generate data quickly and accurately to drive computer simulations. The central idea is to transform a Gaussian vector autoregressive process into the desired multivariate time-series input process that we presume as having a VARTA (Vector-Autoregressive-To-Anything) distribution. We manipulate the correlation structure of the Gaussian vector autoregressive process so that we achieve the desired correlation structure for the simulation input process. For the purpose of computational efficiency, we provide a numerical method, which incorporates a numerical-search procedure and a numerical-integration technique, for solving this correlation-matching problem. Keywords: Computer simulation, vector autoregressive process, vector time series, multivariate input mod- eling, numerical integration. 1 1 Introduction Representing the uncertainty or randomness in a simulated system by an input model is one of the challenging problems in the practical application of computer simulation. There are an abundance of examples, from manufacturing to service applications, where input modeling is critical; e.g., modeling the time to failure for a machining process, the demand per unit time for inventory of a product, or the times between arrivals of calls to a call center. Further, building a large- scale discrete-event stochastic simulation model may require the development of a large number of, possibly multivariate, input models. Development of these models is facilitated by accurate and automated (or nearly automated) input modeling support.
    [Show full text]
  • Fast Estimation of the Median Covariation Matrix with Application to Online Robust Principal Components Analysis
    Fast Estimation of the Median Covariation Matrix with Application to Online Robust Principal Components Analysis Hervé Cardot, Antoine Godichon-Baggioni Institut de Mathématiques de Bourgogne, Université de Bourgogne Franche-Comté, 9, rue Alain Savary, 21078 Dijon, France July 12, 2016 Abstract The geometric median covariation matrix is a robust multivariate indicator of dis- persion which can be extended without any difficulty to functional data. We define estimators, based on recursive algorithms, that can be simply updated at each new observation and are able to deal rapidly with large samples of high dimensional data without being obliged to store all the data in memory. Asymptotic convergence prop- erties of the recursive algorithms are studied under weak conditions. The computation of the principal components can also be performed online and this approach can be useful for online outlier detection. A simulation study clearly shows that this robust indicator is a competitive alternative to minimum covariance determinant when the dimension of the data is small and robust principal components analysis based on projection pursuit and spherical projections for high dimension data. An illustration on a large sample and high dimensional dataset consisting of individual TV audiences measured at a minute scale over a period of 24 hours confirms the interest of consider- ing the robust principal components analysis based on the median covariation matrix. All studied algorithms are available in the R package Gmedian on CRAN. Keywords. Averaging, Functional data, Geometric median, Online algorithms, Online principal components, Recursive robust estimation, Stochastic gradient, Weiszfeld’s algo- arXiv:1504.02852v5 [math.ST] 9 Jul 2016 rithm.
    [Show full text]
  • Characteristics and Statistics of Digital Remote Sensing Imagery (1)
    Characteristics and statistics of digital remote sensing imagery (1) Digital Images: 1 Digital Image • With raster data structure, each image is treated as an array of values of the pixels. • Image data is organized as rows and columns (or lines and pixels) start from the upper left corner of the image. • Each pixel (picture element) is treated as a separate unite. Statistics of Digital Images Help: • Look at the frequency of occurrence of individual brightness values in the image displayed • View individual pixel brightness values at specific locations or within a geographic area; • Compute univariate descriptive statistics to determine if there are unusual anomalies in the image data; and • Compute multivariate statistics to determine the amount of between-band correlation (e.g., to identify redundancy). 2 Statistics of Digital Images It is necessary to calculate fundamental univariate and multivariate statistics of the multispectral remote sensor data. This involves identification and calculation of – maximum and minimum value –the range, mean, standard deviation – between-band variance-covariance matrix – correlation matrix, and – frequencies of brightness values The results of the above can be used to produce histograms. Such statistics provide information necessary for processing and analyzing remote sensing data. A “population” is an infinite or finite set of elements. A “sample” is a subset of the elements taken from a population used to make inferences about certain characteristics of the population. (e.g., training signatures) 3 Large samples drawn randomly from natural populations usually produce a symmetrical frequency distribution. Most values are clustered around the central value, and the frequency of occurrence declines away from this central point.
    [Show full text]
  • Covariance of Cross-Correlations: Towards Efficient Measures for Large-Scale Structure
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by RERO DOC Digital Library Mon. Not. R. Astron. Soc. 400, 851–865 (2009) doi:10.1111/j.1365-2966.2009.15490.x Covariance of cross-correlations: towards efficient measures for large-scale structure Robert E. Smith Institute for Theoretical Physics, University of Zurich, Zurich CH 8037, Switzerland Accepted 2009 August 4. Received 2009 July 17; in original form 2009 June 13 ABSTRACT We study the covariance of the cross-power spectrum of different tracers for the large-scale structure. We develop the counts-in-cells framework for the multitracer approach, and use this to derive expressions for the full non-Gaussian covariance matrix. We show that for the usual autopower statistic, besides the off-diagonal covariance generated through gravitational mode- coupling, the discreteness of the tracers and their associated sampling distribution can generate strong off-diagonal covariance, and that this becomes the dominant source of covariance as spatial frequencies become larger than the fundamental mode of the survey volume. On comparison with the derived expressions for the cross-power covariance, we show that the off-diagonal terms can be suppressed, if one cross-correlates a high tracer-density sample with a low one. Taking the effective estimator efficiency to be proportional to the signal-to-noise ratio (S/N), we show that, to probe clustering as a function of physical properties of the sample, i.e. cluster mass or galaxy luminosity, the cross-power approach can outperform the autopower one by factors of a few.
    [Show full text]
  • Random Processes
    Chapter 6 Random Processes Random Process • A random process is a time-varying function that assigns the outcome of a random experiment to each time instant: X(t). • For a fixed (sample path): a random process is a time varying function, e.g., a signal. – For fixed t: a random process is a random variable. • If one scans all possible outcomes of the underlying random experiment, we shall get an ensemble of signals. • Random Process can be continuous or discrete • Real random process also called stochastic process – Example: Noise source (Noise can often be modeled as a Gaussian random process. An Ensemble of Signals Remember: RV maps Events à Constants RP maps Events à f(t) RP: Discrete and Continuous The set of all possible sample functions {v(t, E i)} is called the ensemble and defines the random process v(t) that describes the noise source. Sample functions of a binary random process. RP Characterization • Random variables x 1 , x 2 , . , x n represent amplitudes of sample functions at t 5 t 1 , t 2 , . , t n . – A random process can, therefore, be viewed as a collection of an infinite number of random variables: RP Characterization – First Order • CDF • PDF • Mean • Mean-Square Statistics of a Random Process RP Characterization – Second Order • The first order does not provide sufficient information as to how rapidly the RP is changing as a function of timeà We use second order estimation RP Characterization – Second Order • The first order does not provide sufficient information as to how rapidly the RP is changing as a function
    [Show full text]
  • On the Use of the Autocorrelation and Covariance Methods for Feedforward Control of Transverse Angle and Position Jitter in Linear Particle Beam Accelerators*
    '^C 7 ON THE USE OF THE AUTOCORRELATION AND COVARIANCE METHODS FOR FEEDFORWARD CONTROL OF TRANSVERSE ANGLE AND POSITION JITTER IN LINEAR PARTICLE BEAM ACCELERATORS* Dean S. Ban- Advanced Photon Source, Argonne National Laboratory, 9700 S. Cass Ave., Argonne, IL 60439 ABSTRACT It is desired to design a predictive feedforward transverse jitter control system to control both angle and position jitter m pulsed linear accelerators. Such a system will increase the accuracy and bandwidth of correction over that of currently available feedback correction systems. Intrapulse correction is performed. An offline process actually "learns" the properties of the jitter, and uses these properties to apply correction to the beam. The correction weights calculated offline are downloaded to a real-time analog correction system between macropulses. Jitter data were taken at the Los Alamos National Laboratory (LANL) Ground Test Accelerator (GTA) telescope experiment at Argonne National Laboratory (ANL). The experiment consisted of the LANL telescope connected to the ANL ZGS proton source and linac. A simulation of the correction system using this data was shown to decrease the average rms jitter by a factor of two over that of a comparable standard feedback correction system. The system also improved the correction bandwidth. INTRODUCTION Figure 1 shows the standard setup for a feedforward transverse jitter control system. Note that one pickup #1 and two kickers are needed to correct beam position jitter, while two pickup #l's and one kicker are needed to correct beam trajectory-angle jitter. pickup #1 kicker pickup #2 Beam Fast loop Slow loop Processor Figure 1. Feedforward Transverse Jitter Control System It is assumed that the beam is fast enough to beat the correction signal (through the fast loop) to the kicker.
    [Show full text]
  • Lecture 4 Multivariate Normal Distribution and Multivariate CLT
    Lecture 4 Multivariate normal distribution and multivariate CLT. T We start with several simple observations. If X = (x1; : : : ; xk) is a k 1 random vector then its expectation is × T EX = (Ex1; : : : ; Exk) and its covariance matrix is Cov(X) = E(X EX)(X EX)T : − − Notice that a covariance matrix is always symmetric Cov(X)T = Cov(X) and nonnegative definite, i.e. for any k 1 vector a, × a T Cov(X)a = Ea T (X EX)(X EX)T a T = E a T (X EX) 2 0: − − j − j � We will often use that for any vector X its squared length can be written as X 2 = XT X: If we multiply a random k 1 vector X by a n k matrix A then the covariancej j of Y = AX is a n n matrix × × × Cov(Y ) = EA(X EX)(X EX)T AT = ACov(X)AT : − − T Multivariate normal distribution. Let us consider a k 1 vector g = (g1; : : : ; gk) of i.i.d. standard normal random variables. The covariance of g is,× obviously, a k k identity × matrix, Cov(g) = I: Given a n k matrix A, the covariance of Ag is a n n matrix × × � := Cov(Ag) = AIAT = AAT : Definition. The distribution of a vector Ag is called a (multivariate) normal distribution with covariance � and is denoted N(0; �): One can also shift this disrtibution, the distribution of Ag + a is called a normal distri­ bution with mean a and covariance � and is denoted N(a; �): There is one potential problem 23 with the above definition - we assume that the distribution depends only on covariance ma­ trix � and does not depend on the construction, i.e.
    [Show full text]