Sequential change-point detection for mixing random sequences under composite hypotheses

The problem of sequential detection of a change-point in the density function of one-dimensional distribution of observations from a mixing random sequence is considered when both before and after a change-point this density function belongs to a certain family of distributions, i.e. in the situation of composite hypotheses. A new quality criterion for change-point detection is proposed. The asymptotic a priori lower bound for this criterion is proved for wide class of methods of change-point detection. An asymptotically optimal method of change-point detection is proposed for which this lower bound is attained asymptotically. In particular, for the case of a simple hypothesis before a change-point, this method coincides with the generalized cumulative sums (CUSUM) method.      

Inference in a model with at most one slope-change point

In this paper the problem of slope-change point in linear regression model is discussed with the help of the theory of Gaussian process. The distribution of the estimators of the change point proposed in this paper can be approximated by the first type of extremal distribution. Based on this fact, the detection and interval estimation of a change-point in various situations are discussed.     

Change-point detection of failure mechanism for electronic devices based on Arrhenius model

In this paper, we consider the abrupt change-point and continuous change-point failure mechanisms in the accelerated life tests of electronic devices and propose corresponding detection methods based on the Weibull distribution and the Arrhenius model. The new methods can handle the detection of failure mechanism change-point with censored data and small samples. Comprehensive simulation studies are conducted to test the performance of the proposed methods and investigate the influences of the change location, the parameters and the censored ratios. Results of the simulation studies show that the proposed methods perform well in terms of type I errors and powers, and the change of parameters has a stronger impact on the performance of the new methods compared to the change location and censored ratios. In the end, we provide two real-world examples. The new class H insulation example shows that the proposed methods can predict the reliability reasonably well by detecting the change point, and the metal oxide semiconductor transistor example shows that the new methods can greatly reduce the experimental time and provide accurate reliability predictions.     

位置参数变点的非参数检验及其渐近性质

本文基于Ｕ－统计量,对于位置参数模型,讨论了位置参数变点的检验问题,给出了检验统计量并研究它的分市的极限性质,证明了检验统计量的极限分布是ｓｕｐ|Ｂ(ｔ)|,其中{Ｂ(ｔ),０＜ｔ＜１}是一个Ｂｒｏｗｎ桥．将此结果应用到了双参数指数分布和Ｗｅｉｂｕｌｌ分布尺度参数变点的检验问题中．     

ARCH过程的均值变点估计

研究ARCH过程的均值变点估计.在较弱的条件下证明了变点估计的一致性,并得到了估计的收敛率;为构造变点的置信区间给出了变点的极限分布.模拟结果表明方法的有效性.     

艾拉姆咖分布参数变点的Bayes估计

研究了艾拉姆咖分布变点估计的非迭代抽样算法(IBF)和MCMC算法.在贝叶斯框架下,选取无信息先验分布,得到关于变点位置的后验分布和各参数的满条件分布,并且详细介绍了IBF算法和MCMC方法的实施步骤.最后进行随机模拟试验,结果表明两种算法都能够有效的估计变点位置,并且IBF算法的计算速度优于MCMC方法.     

Asymptotic Distribution of the Jump Change-Point Estimator

The asymptotic distribution of the change-point estimator in a jump changepoint model is considered.For the jump change-point model Xi =a + θI{[nTo] ＜ i ≤n} + εi,where εi (i =1,…,n) are independent ide...     

An Analytic Expression for the Distribution of the Generalized Shiryaev–Roberts Diffusion

We consider the quickest change-point detection problem where the aim is to detect the onset of a pre-specified drift in “live”-monitored standard Brownian motion; the change-point is assumed unknown (nonrandom). The topic of interest is the distribution of the Generalized Shryaev–Roberts (GSR) detection statistic set up to “sense” the presence of the drift. Specifically, we derive a closed-form formula for the transition probability density function (pdf) of the time-homogeneous Markov diffusion process generated by the GSR statistic when the Brownian motion under surveillance is “drift-free”, i.e., in the pre-change regime; the GSR statistic’s (deterministic) nonnegative headstart is assumed arbitrarily given. The transition pdf formula is found analytically, through direct solution of the respective Kolmogorov forward equation via the Fourier spectral method to achieve separation of the spacial and temporal variables. The obtained result generalizes the well-known formula for the (pre-change) stationary distribution of the GSR statistic: the latter’s stationary distribution is the temporal limit of the distribution sought in this work. To conclude, we exploit the obtained formula numerically and briefly study the pre-change behavior of the GSR statistic versus three factors: (a) drift-shift magnitude, (b) time, and (c) the GSR statistic’s headstart.     

On a multi-channel change-point problem

A continuous change-point problem is studied in which N independent diffusion processes X _j are observed. Each process X _j is associated with a “channel”, each has an unknown piecewise constant drift and the unit diffusion coefficient. All the channels are connected only by a common change-point of drift. As the result, a change-point problem is defined in which the unknown and unidentifiable drift forms a 2N-dimensional nuisance parameter. The asymptotics of the minimax rate in estimating the change-point is studied as N → ∞. This rate is compared with the case of the known drift. This problem is a special case of an open change-point detection problem in the high-dimensional diffusion with nonparametric drift.      

Exploring the latent segmentation space for the assessment of multiple change-point models

This paper addresses the retrospective or off-line multiple change-point detection problem. Multiple change-point models are here viewed as latent structure models and the focus is on inference concerning the latent segmentation space. Methods for exploring the space of possible segmentations of a sequence for a fixed number of change points may be divided into two categories: (i) enumeration of segmentations, (ii) summary of the possible segmentations in change-point or segment profiles. Concerning the first category, a dynamic programming algorithm for computing the top $N$ most probable segmentations is derived. Concerning the second category, a forward-backward dynamic programming algorithm and a smoothing-type forward-backward algorithm for computing two types of change-point and segment profiles are derived. The proposed methods are mainly useful for exploring the segmentation space for successive numbers of change points and provide a set of assessment tools for multiple change-point models that can be applied both in a non-Bayesian and a Bayesian framework. We show using examples that the proposed methods may help to compare alternative multiple change-point models (e.g. Gaussian model with piecewise constant variances or global variance), predict supplementary change points, highlight overestimation of the number of change points and summarize the uncertainty concerning the position of change points.     

Entropy-based Inhomogeneity Detection in Fiber Materials

We study a change-point problem for random fields based on a univariate detection of outliers via the 3σ-rule in order to recognize inhomogeneities in glass fiber reinforced polymers (GFRP). In particular, we focus on GFRP modeled by stochastic fiber processes with high fiber intensity and search for abrupt changes in the direction of the fibers. As a measure of change, the entropy of the directional distribution is locally estimated within a window that scans the region to be analyzed.     

A Bayesian-martingale approach to the general disorder problem

We consider a Bayesian-martingale approach to the general change-point detection problem. In our setting the change-point represents a random time of bifurcation of two probability measures given on the space of right-continuous functions. We derive a reflecting backward stochastic differential equation (RBSDE) for the value process related to the disorder problem and show that in classical cases of the Wiener and Poisson disorder problems this RBSDE is equivalent to free-boundary problems for parabolic differential and differential–difference operators respectively.     

Poisson source localization on the plane: change-point case

We present a detection problem where several spatially distributed sensors observe Poisson signals emitted from a single radioactive source of unknown position. The measurements at each sensor are modeled by independent inhomogeneous Poisson processes. A method based on Bayesian change-point estimation is proposed to identify the location of the source’s coordinates. The asymptotic behavior of the Bayesian estimator is studied. In particular, the consistency and the asymptotic efficiency of the estimator are analyzed. The limit distribution and the convergence of the moments are also described. The similar statistical model could be used in GPS localization problems.

     

至多一个分布变点的非参数检验及其渐近性质

本文研究了连续分布函数变点的假设检验问题,通过秩统计量和次序统计量方法,得到了相应的检验统计量及其渐近性质.     

Estimation of a change-point in the mean function of functional data

The paper develops a comprehensive asymptotic theory for the estimation of a change-point in the mean function of functional observations. We consider both the case of a constant change size, and the case of a change whose size approaches zero, as the sample size tends to infinity. We show how the limit distribution of a suitably defined change-point estimator depends on the size and location of the change. The theoretical insights are confirmed by a simulation study which illustrates the behavior of the estimator in finite samples.     

An Efficient Algorithm for Bayesian Nearest Neighbours

K-Nearest Neighbours (k-NN) is a popular classification and regression algorithm, yet one of its main limitations is the difficulty in choosing the number of neighbours. We present a Bayesian algorithm to compute the posterior probability distribution for k given a target point within a data-set, efficiently and without the use of Markov Chain Monte Carlo (MCMC) methods or simulation—alongside an exact solution for distributions within the exponential family. The central idea is that data points around our target are generated by the same probability distribution, extending outwards over the appropriate, though unknown, number of neighbours. Once the data is projected onto a distance metric of choice, we can transform the choice of k into a change-point detection problem, for which there is an efficient solution: we recursively compute the probability of the last change-point as we move towards our target, and thus de facto compute the posterior probability distribution over k. Applying this approach to both a classification and a regression UCI data-sets, we compare favourably and, most importantly, by removing the need for simulation, we are able to compute the posterior probability of k exactly and rapidly. As an example, the computational time for the Ripley data-set is a few milliseconds compared to a few hours when using a MCMC approach.

     

Testing for the change of the mean-reverting parameter of an autoregressive model with stationary Gaussian noise

The likelihood ratio test for a change in the mean-reverting parameter of a first order autoregressive model with stationary Gaussian noise is considered. The test statistic converges in distribution to the Gumbel extreme value distribution under the null hypothesis of no change-point for a large class of covariance structures including long-memory processes as the fractional Gaussian noise.

     

A change-point model for the r-largest order statistics with applications to environmental and financial data

This study makes a new contribution to extreme value theory by proposing a change-point model of the distribution of the r-larger order statistics. In some situations, using only the maxima of grouped data results in a small sample size that may require a larger dataset. In this sense, using the joint distribution of the r-largest order statistics provides more information and, consequently, better estimators. We perform a comprehensive simulation to show the advantage of this method over other competitive models that approach the change-point model in extremes. Finally, the proposed model is fitted to river quota data (environmental data) and NASDAQ daily returns data (financial data) to demonstrate its potential for practical application.     

基于贝叶斯推断的TAR模型的门限非线性检验

在正态分布的假定下,变点问题按照均值和方差的变化有四种情形.本文把TAR模型门限非线性的检验问题,看作是对应均值变化,方差不变情形下的变点问题.然后利用可逆跳马尔可夫蒙特卡罗模拟(RJMCMC)方法计算两个比较模型(AR和TAR模型)的后验概率.后验概率的结果支持TAR模型表明门限非线性的存在.模拟实验的结果说明基于贝叶斯推断的检验方法可以很好的区分AR和TAR模型.     

A note on testing for constant hazard against a change-point alternative

Summary The problem of testing for constant hazard against a change-point alternative is considered. It is shown that this problem is related to another one in quality control. Based on this relationship, a test is proposed. The main advantages of this test are its computational simplicity and the ready availability of small and large sample distribution theory.