《Journal of nonparametric statistics》2012,24(4):935-949

Smooth nonparametric estimators based on a kernel method are proposed for cumulative distribution functions (CDFs) and quantiles of circular data. A sound motivation for this is that although for euclidean data similar estimators have been widely studied, for circular data nothing similar seems to exist; albeit, remarkably, in the circular-setting local methods are implemented more easily because of the absence of boundaries on the circle. The only alternative to our method seems to be the empirical CDF, that does not take into account circularity of data when the estimate is near the cut-point, as our local method naturally does. The definition of circular CDF is different from its euclidean counterpart in many respects, and this will give rise to estimators exhibiting some ‘unusual’ features such as, for example, global efficiency measures containing a location parameter and a covariance term. Simulations along with real data case studies illustrate the findings.  相似文献   

    

《Journal of nonparametric statistics》2012,24(1):91-101

The asymptotic behavior of nonparametric estimates of the regression and autoregression function based on local L-estimates is considered. The results are stated separately for score functions leading to robust estimates and for score functions whose support is the whole [0,1] interval. Weaker assumptions are required for the robust estimates.  相似文献   

    

《Journal of nonparametric statistics》2012,24(4):323-335

When the generalized Wilcoxon test is applied to the mouse leukemia data in Kalbflesch and Prentice (1980), it fails to detect a difference between the lifetime distributions of the male and the female mice which died from thymic leukemia. This may be a result of the insensitivity of the test in detecting a distributional difference in location and scale simultaneously. In this article, a simple quantile-based approach to the two-sample problem under a location-scale model is studied. Large sample inference on the location and the scale parameters is discussed. Subject to random right censoring, a preliminary test is proposed to detect a scale change in the location-scale model against the pure location model. The mouse leukemia data are used to illustrate the proposed approach.  相似文献   

    

《Journal of nonparametric statistics》2012,24(2):149-163

We establish a weak invariance principle for certain functionals of the regressogram estimator for regression or autoregression models where the data are strongly mixing. These functionals are constructed by cumulating the local discrepancies between the regressogram estimator and the corresponding regression function. As a byproduct, we obtain the limiting distribution of these functionals. Since the limiting process turns out to be a tractable time-changed Wiener process, we can derive from our results a family of possible nonparametric goodness-of-fit tests for the restriction to any compact interval of the regression or autoregression function. We then focus on a specially interesting test within this family. Using our preceding results, we provide estimates for the asymptotic behavior of the power of this test against both fixed and local alternatives.  相似文献   

    

《Journal of nonparametric statistics》2012,24(2):265-282

This paper develops recursive kernel estimators for the probability density and the regression function of nonlinear and nonstationary time series. The resulting method is characterized by two smoothing coefficients (the bandwidth and the discounting rate of observations) that may be selected with a prediction error criterion. Statistical properties are investigated under a null hypothesis of stationarity and asymptotic elimination of the discounting. Simulation experiments on complex processes show the ability of the method in estimating time-varying nonlinear regression functions.  相似文献   

    

《Journal of nonparametric statistics》2012,24(1-2):203-222

We discuss the interplay between local M -smoothers, Bayes smoothers and some nonlinear filters for edge-preserving signal reconstruction. We prove that all smoothers in question are nonlinear filters in a precise sense and characterize their fixed points. Then a Potts model is adopted for segmentation. For 1-d signals, an exact algorithm for the computation of maximum posterior modes is derived and applied to a phantom and to 1-d fMRI-data.  相似文献   

    

《Journal of nonparametric statistics》2012,24(1):85-97

A new characterisation of the exponential distribution in a wide class of new better than used in pth quantile (NBUp) lifetime distributions is presented. This leads to new classes of scale-free goodness-of-fit tests for exponentiality against NBUp alternatives. The limiting distributions of the test statistics under the null and alternative hypotheses are derived and the tests are shown to be consistent against NBUp alternatives. Pitman efficacies are calculated and a limited Monte Carlo study is conducted to compare the tests with regard to power for small and moderate sample sizes against a range of alternative distributions. On the basis of overall good performance and ease of computation, a member of the class of test statistics which is based on the sample Winsorised mean is recommended as a scale-free goodness-of-fit test for the exponential distribution.  相似文献   

    

《Journal of nonparametric statistics》2012,24(8):1039-1051

  相似文献   

    

《Journal of nonparametric statistics》2012,24(1):59-65

In this paper, we investigate the Bahadur representation of sample quantile based on negatively orthant dependent sequence, which is weaker than negatively associated sequence. Our results extend and improve the results of Ling [(2008), ‘The Bahadur Representation for Sample Quantiles Under Negatively Associated Sequence’, Statistics & Probability Letters, 78, 2660–2663].  相似文献   

    

《Journal of nonparametric statistics》2012,24(4):993-1006

In this paper, we empirically investigate the robustness of random forests for regression problems. We also investigate the performance of six variations of the original random forest method, all aimed at improving robustness. These variations are based on three main ideas: (1) robustify the aggregation method, (2) robustify the splitting criterion and (3) taking a robust transformation of the response. More precisely, with the first idea, we use the median (or weighted median), instead of the mean, to combine the predictions from the individual trees. With the second idea, we use least-absolute deviations from the median, instead of least-squares, as splitting criterion. With the third idea, we build the trees using the ranks of the response instead of the original values. The competing methods are compared via a simulation study with artificial data using two different types of contaminations and also with 13 real data sets. Our results show that all three ideas improve the robustness of the original random forest algorithm. However, a robust aggregation of the individual trees is generally more profitable than a robust splitting criterion.  相似文献