Search

A nonparametric procedure to estimate the conditional probability that a nonstationary geostatistical process exceeds a certain threshold value is proposed. The method consists of a bootstrap algorithm that combines conditional simulation techniques with nonparametric estimations of the trend and the variability. The nonparametric local linear estimator, considering a bandwidth matrix selected by a method that takes the spatial dependence into account, is used to estimate the trend. The variability is modeled estimating the conditional variance and the variogram from corrected residuals to avoid the biasses. The proposed method allows to obtain estimates of the conditional exceedance risk in non-observed spatial locations. The performance of the approach is analyzed by simulation and illustrated with the application to a real data set of precipitations in the USA.Supplementary materials accompanying this paper appear on-line. [ABSTRACT FROM AUTHOR]

The aim of this study is to statistically identify and distinguish wood samples corresponding to different areas of annual rings in trees of temperate regions, using the corresponding thermogravimetric (TG) and their first TG derivative (DTG) curves, and specifically to verify whether late and early wood chestnut samples are different with statistical significance, taking into account their TG and DTG curves. These significant differences are sought by applying statistical procedures based on functional data analysis (FDA), such as the functional ANOVA and the FDA classification methods. Each TG curve is firstly smoothed using the local polynomial regression estimator, and its first derivative is estimated. Then, functional ANOVA based on random projections (RP) is used to identify significant differences between TG or DTG curves of early and late wood samples. In order to know the extent of the differences between early and late wood samples, they are discriminated (and the correct classification proportion obtained) by employing a kernel nonparametric functional data analysis technique, based on the Bayes' rule, as well as functional generalized linear models and functional generalized additive models, allowing to classify materials using more than one type of thermal curves simultaneously. The results are compared with those obtained using some classical multivariate supervised classification methods: linear discriminant analysis, naive Bayes (NBC) and quadratic classification (QDA). The partial least squares (PLS) dimension reduction procedure was previously applied to the TG curves in order to employ these multivariate methods. The application of RP ANOVA shows significant differences between late and early wood regarding mass loss and mass loss rate under combustion. The use of PLS multivariate methods or FDA classification approaches applied to the TG and DTG curves allows to distinguish very accurately between late and early wood. The proposed method could be applied to other species to identify thermooxidative differences, combined with other experimental methods to find their chemical and physical causes. [ABSTRACT FROM AUTHOR]

Interval-grouped data are defined, in general, when the event of interest cannot be directly observed and it is only known to have been occurred within an interval. In this framework, a nonparametric kernel density estimator is proposed and studied. The approach is based on the classical Parzen–Rosenblatt estimator and on the generalisation of the binned kernel density estimator. The asymptotic bias and variance of the proposed estimator are derived under usual assumptions, and the effect of using non-equally spaced grouped data is analysed. Additionally, a plug-in bandwidth selector is proposed. Through a comprehensive simulation study, the behaviour of both the estimator and the plug-in bandwidth selector considering different scenarios of data grouping is shown. An application to real data confirms the simulation results, revealing the good performance of the estimator whenever data are not heavily grouped. [ABSTRACT FROM AUTHOR]

In this paper, nonparametric curve estimation methods are applied to analyze time series of wind speeds, focusing on the extreme events exceeding a chosen threshold. Classical parametric statistical approaches in this context consist in fitting a generalized Pareto distribution (GPD) to the tail of the empirical cumulative distribution, using maximum likelihood or the method of the moments to estimate the parameters of this distribution. Additionally, confidence intervals are usually computed to assess the uncertainty of the estimates. Nonparametric methods to estimate directly some quantities of interest, such as the probability of exceedance, the quantiles or return levels, or the return periods, are proposed. Moreover, bootstrap techniques are used to develop pointwise and simultaneous confidence intervals for these functions. The proposed models are applied to wind speed data in the Gulf Coast of US, comparing the results with those using the GPD approach, by means of a split-sample test. Results show that nonparametric methods are competitive with respect to the standard GPD approximations. The study is completed generating synthetic data sets and comparing the behavior of the parametric and the nonparametric estimates in this framework. [ABSTRACT FROM AUTHOR]

Abstract: Nonparametric techniques are frequently used in seismology studies to fit statistical models to earthquake data, as an alternative to classical parametric methods. In this paper, we apply nonparametric techniques to study some area-characteristic seismic hazard parameters (such as the mean return period or the expected number of earthquakes) in the northwest of the Iberian Peninsula, using a complete data catalog. Moreover, bootstrap techniques are used to develop simultaneous confidence bands for these functions. The previous nonparametric approaches are adapted to a spatial framework, obtaining maps of the probabilities of exceedance of the annual maximum magnitude in the zone of study. This procedure suitably describes the features of the area studied, being consistent with previous geological researches. [Copyright &y& Elsevier]

Abstract: Controlling emissions of air pollutants and establishing air quality objectives to improve and protect ambient air quality are very important tasks of Governments. Ozone (O3), one of those pollutants of concern, is not emitted directly into the atmosphere, but is a secondary pollutant produced by reaction between nitrogen dioxide (NO2), hydrocarbons and sunlight. High levels of ozone can produce harmful effects on human health and the environment in general. Therefore, the study of extreme values of ozone represents an important topic of research in environmental problems. Classical extreme value theory has been usually used in air-pollution studies. It consists of fitting a parametric generalized extreme value (GEV) distribution to a data set of extreme values and using the estimated distribution to compute quantities like the probability of exceedance, the quantiles, the return levels or the mean return periods. In this paper, we propose nonparametric methods to estimate those quantities. Additionally, nonparametric estimators of the trends of very high values of ozone are proposed. The nonparametric estimators are applied to real samples of maximum ozone values obtained from several monitoring stations belonging to the Automatic Urban and Rural Network (AURN) from the UK. Results show that nonparametric estimators work satisfactorily, generally outperforming the behaviour of classical parametric estimators. [ABSTRACT FROM AUTHOR]

In this work, a fully nonparametric geostatistical approach to estimate threshold exceeding probabilities is proposed. To estimate the large-scale variability (spatial trend) of the process, the nonparametric local linear regression estimator, with the bandwidth selected by a method that takes the spatial dependence into account, is used. A bias-corrected nonparametric estimator of the variogram, obtained from the nonparametric residuals, is proposed to estimate the small-scale variability. Finally, a bootstrap algorithm is designed to estimate the unconditional probabilities of exceeding a threshold value at any location. The behavior of this approach is evaluated through simulation and with an application to a real data set. [ABSTRACT FROM AUTHOR]