Search results for "Order statistic"
showing 10 items of 29 documents
Adaptive algorithms robust to impulsive noise with low computational cost using order statistic
2009
Abstract In this paper a family of adaptive algorithms robust to impulsive noise and with low computational cost are presented. Unlike other approaches, no cost functions or filtering of the gradient are considered in order to update the filter coefficients. Its initial basis is the basic LMS algorithm and its sign-error variant. The proposed algorithms can be considered as some sign-error variants of the LMS algorithm. The algorithms are successfully tested in terms of accuracy and convergence in a standard system identification simulation in which an impulsive noise is present. Simulations show that they improve the performance of LMS variants that are robust to impulsive noise.
The fundamental theory of optimal "Anti-Bayesian" parametric pattern classification using order statistics criteria
2013
Author's version of an article in the journal: Pattern Recognition. Also available from the publisher at: http://dx.doi.org/10.1016/j.patcog.2012.07.004 The gold standard for a classifier is the condition of optimality attained by the Bayesian classifier. Within a Bayesian paradigm, if we are allowed to compare the testing sample with only a single point in the feature space from each class, the optimal Bayesian strategy would be to achieve this based on the (Mahalanobis) distance from the corresponding means. The reader should observe that, in this context, the mean, in one sense, is the most central point in the respective distribution. In this paper, we shall show that we can obtain opti…
Higher order statistics of the response of MDOF linear systems excited by linearly parametric white noises and external excitations
1997
The aim of this paper is the evaluation of higher order statistics of the response of linear systems subjected to external excitations and to linearly parametric white noise. The external excitations considered are deterministic or filtered white noise processes. The procedure implies the knowledge of the transition matrix connected to the linear system; this, however, has already been evaluated for obtaining the statistics at single times. The method, which avoids making further integrations for the evaluation of the higher order statistics, is very advantageous from a computational point of view.
Higher order statistics of the response of MDOF linear systems under polynomials of filtered normal white noises
1997
This paper exploits the work presented in the companion paper in order to evaluate the higher order statistics of the response of linear systems excited by polynomials of filtered normal processes. In fact, by means of a variable transformation, the original system is replaced by a linear one excited by external and linearly parametric white noise excitations. The transition matrix of the new enlarged system is obtained simply once the transition matrices of the original system and of the filter are evaluated. The method is then applied in order to evaluate the higher order statistics of the approximate response of nonlinear systems to which the pseudo-force method is applied.
Higher order statistics of the response of linear systems excited by polynomials of filtered Poisson pulses
1999
The higher order statistics of the response of linear systems excited by polynomials of filtered Poisson pulses are evaluated by means of knowledge of the first order statistics and without any further integration. This is made possible by a coordinate transformation which replaces the original system by a quasi-linear one with parametric Poisson delta-correlated input; and, for these systems, a simple relationship between first order and higher order statistics is found in which the transition matrix of the dynamical new system, incremented by the correction terms necessary to apply the Ito calculus, appears.
Weibull Prediction Limits for a Future Number of Failures Under Parametric Uncertainty
2012
In this paper, we present an accurate procedure, called “within-sample prediction of order statistics,” to obtain prediction limits for the number of failures that will be observed in a future inspection of a sample of units, based only on the results of the first in-service inspection of the same sample. The failure-time of such units is modeled with a two-parameter Weibull distribution indexed by scale and shape parameters β and δ, respectively. It will be noted that in the literature only the case is considered when the scale parameter β is unknown, but the shape parameter δ is known. As a rule, in practice the Weibull shape parameter δ is not known. Instead it is estimated subjectively …
Cluster size distributions in particle systems with asymmetric dynamics
2001
We present exact and asymptotic results for clusters in the one-dimensional totally asymmetric exclusion process (TASEP) with two different dynamics. The expected length of the largest cluster is shown to diverge logarithmically with increasing system size for ordinary TASEP dynamics and as a logarithm divided by a double logarithm for generalized dynamics, where the hopping probability of a particle depends on the size of the cluster it belongs to. The connection with the asymptotic theory of extreme order statistics is discussed in detail. We also consider a related model of interface growth, where the deposited particles are allowed to relax to the local gravitational minimum.
Diagnostics for nonparametric estimation in space-time seismic processes
2010
In this paper we propose a nonparametric method, based on locally variable bandwidths kernel estimators, to describe the space-time variation of seismic activity of a region of Southern California. The flexible estimation approach is introduced together with a diagnostic method for space-time point process, based on the interpretation of some second-order statistics, to analyze the dependence structure of observed data and suggest directions for fit improvement. In this paper we review a diagnostic method for space-time point processes based on the interpretation of the transformed version of some second-order statistics. The method is useful to analyze dependence structures of observed dat…
Successive Reduction of Arms in Multi-Armed Bandits
2011
The relevance of the multi-armed bandit problem has risen in the past few years with the need for online optimization techniques in Internet systems, such as online advertisement and news article recommendation. At the same time, these applications reveal that state-of-the-art solution schemes do not scale well with the number of bandit arms. In this paper, we present two types of Successive Reduction (SR) strategies - 1) Successive Reduction Hoeffding (SRH) and 2) Successive Reduction Order Statistics (SRO). Both use an Order Statistics based Thompson Sampling method for arm selection, and then successively eliminate bandit arms from consideration based on a confidence threshold. While SRH…
Kernel Spectral Angle Mapper
2016
This communication introduces a very simple generalization of the familiar spectral angle mapper (SAM) distance. SAM is perhaps the most widely used distance in chemometrics, hyperspectral imaging, and remote sensing applications. We show that a nonlinear version of SAM can be readily obtained by measuring the angle between pairs of vectors in a reproducing kernel Hilbert spaces. The kernel SAM generalizes the angle measure to higher-order statistics, it is a valid reproducing kernel, it is universal, and it has consistent geometrical properties that permit deriving a metric easily. We illustrate its performance in a target detection problem using very high resolution imagery. Excellent re…