We study the problem of detecting a change in the mean of one-dimensional Gaussian process data. This problem is investigated in the setting of increasing domain (customarily employed in time series analysis) and in the setting of fixed domain (typically arising in spatial data analysis). We propose a detection method based on the generalized likelihood ratio test (GLRT), and show that our method achieves nearly asymptotically optimal rate in the minimax sense, in both settings. The salient feature of the proposed method is that it exploits in an efficient way the data dependence captured by the Gaussian process covariance structure. When the covariance is not known, we propose the plug-in GLRT method and derive conditions under which the method remains asymptotically near optimal. By contrast, the standard CUSUM method, which does not account for the covariance structure, is shown to be asymptotically optimal only in the increasing domain. Our algorithms and accompanying theory are applicable to a wide variety of covariance structures, including the Matern class, the powered exponential class, and others. The plug-in GLRT method is shown to perform well for maximum likelihood estimators with a dense covariance matrix.
Abstract. The computational and/or communication constraints associated with processing large-scale data sets using support vector machines (SVM) in contexts such as distributed networking systems are often prohibitively high, resulting in practitioners of SVM learning algorithms having to apply the algorithm on approximate versions of the kernel matrix induced by a certain degree of data reduction. In this paper, we study the tradeoffs between data reduction and the loss in an algorithm's classification performance. We introduce and analyze a consistent estimator of the SVM's achieved classification error, and then derive approximate upper bounds on the perturbation on our estimator. The bound is shown to be empirically tight in a wide range of domains, making it practical for the practitioner to determine the amount of data reduction given a permissible loss in the classification performance.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.