Log-modified Weibull regression models with censored data: Sensitivity and residual analysis

This paper proposes a regression model considering the modified Weibull distribution. This distribution can be used to model bathtub-shaped failure rate functions. Assuming censored data, we consider maximum likelihood and Jackknife estimators for the parameters of the model. We derive the appropriate matrices for assessing local influence on the parameter estimates under different perturbation schemes and we also present some ways to perform global influence. Besides, for different parameter settings, sample sizes and censoring percentages, various simulations are performed and the empirical distribution of the modified deviance residual is displayed and compared with the standard normal distribution. These studies suggest that the residual analysis usually performed in normal linear regression models can be straightforwardly extended for a martingale-type residual in log-modified Weibull regression models with censored data. Finally, we analyze a real data set under log-modified Weibull regression models. A diagnostic analysis and a model checking based on the modified deviance residual are performed to select appropriate models. (c) 2008 Elsevier B.V. All rights reserved.

Regression models for grouped survival data: Estimation and sensitivity analysis

In this study, regression models are evaluated for grouped survival data when the effect of censoring time is considered in the model and the regression structure is modeled through four link functions. The methodology for grouped survival data is based on life tables, and the times are grouped in k intervals so that ties are eliminated. Thus, the data modeling is performed by considering the discrete models of lifetime regression. The model parameters are estimated by using the maximum likelihood and jackknife methods. To detect influential observations in the proposed models, diagnostic measures based on case deletion, which are denominated global influence, and influence measures based on small perturbations in the data or in the model, referred to as local influence, are used. In addition to those measures, the local influence and the total influential estimate are also employed. Various simulation studies are performed and compared to the performance of the four link functions of the regression models for grouped survival data for different parameter settings, sample sizes and numbers of intervals. Finally, a data set is analyzed by using the proposed regression models. (C) 2010 Elsevier B.V. All rights reserved.

A Log-Linear Regression Model for the Beta-Weibull Distribution

We introduce the log-beta Weibull regression model based on the beta Weibull distribution (Famoye et al., 2005; Lee et al., 2007). We derive expansions for the moment generating function which do not depend on complicated functions. The new regression model represents a parametric family of models that includes as sub-models several widely known regression models that can be applied to censored survival data. We employ a frequentist analysis, a jackknife estimator, and a parametric bootstrap for the parameters of the proposed model. We derive the appropriate matrices for assessing local influences on the parameter estimates under different perturbation schemes and present some ways to assess global influences. Further, for different parameter settings, sample sizes, and censoring percentages, several simulations are performed. In addition, the empirical distribution of some modified residuals are displayed and compared with the standard normal distribution. These studies suggest that the residual analysis usually performed in normal linear regression models can be extended to a modified deviance residual in the proposed regression model applied to censored data. We define martingale and deviance residuals to evaluate the model assumptions. The extended regression model is very useful for the analysis of real data and could give more realistic fits than other special regression models.

Political Parties and Governors as Determinants of Legislative Behavior in Brazil`s Chamber of Deputies, 1988-2006

This article examines the relative importance of regional and national forces in shaping the behavior of Brazilian legislators at the national level. A widely held view is that national legislators respond to state pressures in making decisions, rather than pressures from the national government. Governors not only can influence national debates but also can determine outcomes by exerting control over their states` legislative delegations. This article examines a dataset of all roll-call votes in the Chamber of Deputies between 1989 and 2006 to isolate and evaluate the impact of local pressures on legislative voting. Spanning the terms of five presidents and five different congresses, the data show that the local influence is weaker than the national on the voting decisions of individual legislators and the voting cohesion of state delegations. Alternative institutional resources allow the central government to counteract the centrifugal pressures of federalism and other institutional influences.

Diagnostic techniques applied in geostatistics for agricultural data analysis

The structural modeling of spatial dependence, using a geostatistical approach, is an indispensable tool to determine parameters that define this structure, applied on interpolation of values at unsampled points by kriging techniques. However, the estimation of parameters can be greatly affected by the presence of atypical observations in sampled data. The purpose of this study was to use diagnostic techniques in Gaussian spatial linear models in geostatistics to evaluate the sensitivity of maximum likelihood and restrict maximum likelihood estimators to small perturbations in these data. For this purpose, studies with simulated and experimental data were conducted. Results with simulated data showed that the diagnostic techniques were efficient to identify the perturbation in data. The results with real data indicated that atypical values among the sampled data may have a strong influence on thematic maps, thus changing the spatial dependence structure. The application of diagnostic techniques should be part of any geostatistical analysis, to ensure a better quality of the information from thematic maps.

Circum-Alpine marine circulation and paleoenvironmental conditions during the Miocene

Trace element and isotopic compositions of marine fossils and sediment were analyzed from several Miocene deposits in the circum-Alpine region in order to reconstruct the paleoceanographic and paleoclimatic changes related to sea level changes, basin evolution and Alpine orogeny. To the north and the east the Alps were border by an epicontinental sea, the Paratethys, while to the south the Mediterranean surrounded the uplifting mountains during the Miocene. The thesis mainly focused on sediments and fossils sampled from Miocene beds of these two oceanic provinces. The north Alpine Molasse, the Vienna and Pannonian Basins were located in the Western and Central Paratethys. O-isotope compositions of well-preserved phosphatic fossils in these sediments support deposition under sub-tropical to warm-temperate climate with water temperatures between 14 to 28 °C for the Miocene. δ18O values of fossil shark teeth from different horizons vary similarly to those of the global trend until the end of the Badenian, however the δ18O values show wider range, which indicates local effects iii the sub-basins. The trend of 87Sr/86Sr in the samples roughly agrees with an open ocean environment for the Miocene. Yet a number of samples deviate from typical open ocean compositions with higher ratios suggesting modification of seawater by local and old terrestrial sources. In contrast, two exceptional teeth from the locality of La Moliere have extremely low δ18O values and low 87Sr/86Sr. However, the REE patterns of their enameloid are similar to those of teeth having O and Sr isotopic compositions typical of a marine setting at this site. Collectively, this suggests that the two teeth formed while the sharks frequented a freshwater environment with very low 18O-content and 87Sr/86Sr controlled by Mesozoic calcareous rocks. This is consistent with a paleogeography of high-elevation (~2300m) Miocene Alps adjacent to a marginal sea. The local effects are also reflected in the εNd values of the Paratethyan fossils, which is compatible with input from ancient crystalline rocks and Mesozoic sediments, while other samples with elevated εNd values indicate an influence of Neogene volcanism on the water budget. Excluding samples whose isotopic compositions reflect a local influence on the water column, an average εNd value of -7.9 ± 0.5 may be inferred for the Paratethys seawater. This value is indistinguishable from the Miocene value of the Indian Ocean, supporting a dominant role of ludo-Pacific water masses in the Paratethys. Regarding the Mediterranean, stable C-and O-isotope compositions of benthic and planktonic foraminifera from the Umbria-Marche region (UMC) have an offset typical for their habitats and the changes in composition mimic global changes, suggesting that the regional conditions of climate and the carbon cycle were controlled by global changes. The radiogenic isotope compositions of the fossil assemblages allow for distinction of periods. From 25 to 19 Ma, high εNd values and low 87Sr/86Sr of sediments and fossils support intense tectonism and volcanism, related to the opening of the western Mediterranean. Between 19 and 13 Ma the Mediterranean has εNd values that are largely controlled by incursion of Indian Ocean water. Brief periods of local hinterland control on seawater compositions are indicated by spikes in the εNd record, coinciding with volcanic events and a short sea-level decrease at about 15.2 Ma. Lower 87Sr/86Sr compared to the open ocean is compatible with rapid uplift of the hinterland and intense influx of Sr from Mesozoic carbonates of the western Apennines, while higher 87Sr/86Sr for other sites indicates erosion of old crustal silicate rocks. Finally, from 13 to 7 Ma the fossils have 87Sr/86Sr similar to those of Miocene seawater and their εNd values indicates fluctuating influence of Atlantic, and Indian Ocean or Paratethys sources of seawater entering the Mediterranean, driven by global sealevel changes and local tectonism. RÉSUMÉ DE LA THÈSE Les compositions en éléments traces et isotopiques de fossiles marins et de sédiments on été analysées à partir de nombreux dépôts marins dans la région circum Alpine dans le but de reconstruire les changements paléocéanographiques et paléoclimatiques liés aux changements du niveau marin, à l'évolution en bassins et à l'orogénie alpine. Au nord et à l'est des Alpes, une mer épicontinentale appelée Paratéthys s'est ouverte, alors que plus au sud la mer Méditerranée bordait au Miocène les Alpes naissantes. Le but de cette recherche est de se concentrer sur les sédiments et les fossiles provenant des couches du Miocènes de ces deux provinces marines. Les bassins de la Molasse Alpine du nord, de Vienne et Pannonien étaient situés au niveau de la Paratéthys Occidentale et Centrale. Les compositions isotopiques de l'oxygène de fossiles phosphatés bien préservés dans ces sédiments étayent la théorie d'un dépôt sous un climat subtropical à tempéré chaud avec des températures entre 14 et 28°C pendant le Miocène. Les valeurs δ18O des fossiles sont similaires à la tendance globale jusqu'à la fin du Badénien. Cependant les larges fluctuations en δ18O indiquent des effets locaux au niveau des sous bassins. En outre, deux dents de requin exceptionnelles présentent des valeurs extrêmement basses de δ18O. Ces données suggèrent que ces deux dents se sont formées alors que les requins fréquentaient un environnement d'eau douce avec de faibles valeurs de 18O. Le calcul de la composition isotopique de l'oxygène de cette eau douce permet d'obtenir une estimation de la paléoélévatian moyenne des Alpes du Miocène (~2300m). La tendance 87Sr/86Sr pour ces échantillons concorde approximativement avec un environnement d'océan ouvert au cours du Miocène. Toutefois un nombre d'échantillons dévie des compositions d'océan ouvert typiques, avec des rapports élevés suggérant des modifications de l'eau de mer par des sources locales et terrestres. Les effets locaux sont aussi reflétés au niveau des valeurs en εNd des fossiles paratéthysiens. Ceci est cohérent avec un apport d'anciennes roches cristallines et de sédiments mésozoïques, tandis que d'autres échantillons avec des valeurs hautes de εNd indiquent une influence d'un volcanisme néogène dans le budget marin. En excluant les échantillons dont les compositions isotopiques confirment une influence locale, une valeur moyenne de εNd de 7.9 t 0.5 peut être déduite pour l'eau de la Parathétys. Cette valeur est semblable à la valeur correspondant à l'Océan Indien durant le Miocène, confirmant un rôle dominant de cet océan dans la Paratéthys. Au niveau de la Méditerranée, les compositions en isotopes stables du Carbone et de l'Oxygène de foraminifères planctoniques et benthique de la région Umbria-Marche présentent un offset typique à leurs habitats. De plus les changements dans leurs compositions suivent les changements globaux, suggérant ainsi que les conditions climatiques régionales et le cycle du carbone étaient contrôlés par des phénomènes globaux. La composition en isotopes radiogéniques d'assemblages fossiles permet une reconnaissance sur trois périodes distinctes. De 25 à 19 millions d'années (Ma), des valeurs élevées de εNd et un faible rapport 87Sr/86Sr dans les sédiments soutiennent l'idée d'une activité tectonique et volcanique intense, liée à l'ouverture de la Méditerranée occidentale. Entre 19 et 13 Ma, la Méditerranée montre des valeurs de εNd qui sont largement contrôlées par une incursion d'eau provenant de l'Océan Indien. En effet, aux alentours de 15,2 Ma, des pics dans l'enregistrement des valeurs de εNd, coïncidant avec des événements volcaniques et de brèves diminutions du niveau marin. Enfin, de 13 à 7 Ma, les fossiles ont des rapports ß7Sr/8fiSr similaires à ceux de l'eau de mer au Miocène. Leurs valeurs de εNd indiquent une influence changeante de l'océan Atlantique, et de l'océan Indien ou des sources d'eau de merde la Parathétys qui entrent dans les bassins méditerranéens. Ce changement est guidé par des modifications globales du niveau marin et par la tectonique locale. RÉSUMÉ DE LA THÈSE (POUR LE GRAND PUBLIC) Les analyses des compositions en éléments traces et isotopiques des fossiles marins sont un outil très utile pour reconstruire les conditions océaniques et climatiques anciennes. Ce travail de thèse se concentre sur les sédiments déposés dans un environnement marin proches des Alpes au cours du Miocène, entre 23 et 7 millions d'années (Ma). Cette période est caractérisée par une tectonique alpine active, ainsi que par des changements climatiques et océanographiques globaux importants. Dans le but de tracer ces changements, les compositions isotopiques du Strontium, du Néodyme, de l'Oxygène et du Carbone ont été analysées dans des fossiles bien préservés ainsi que les sédiments contemporains. Les échantillons proviennent de deux provinces océaniques distinctes, la première est la Mer Méditerranée, et l'autre est une mer épicontinentale appelée Parathétys, qui existait au nord et à l'est des Alpes durant le Miocène. Au niveau de la Parathétys Occidentale et Orientale, les compositions isotopiques d'oxygène de dents de requins confirment un dépôt sous un climat subtropical à tempéré chaud avec des températures d'eau entre 14 et 28°C au Miocène. En outre, deux dents de requins exceptionnelles ont enregistré des compositions isotopiques d'oxygène extrêmement basses. Cela suggère que ces deux dents se sont formées alors que les requins entraient dans un système d'eau douce. Le calcul de la composition isotopique de l'oxygène de cette eau douce permet d'obtenir une estimation de la paléoélévation des Alpes au Miocène qui est aussi élevée que celle d'aujourd'hui. La tendance isotopique du Strontium pour ces échantillons concorde approximativement avec un environnement d'océan ouvert. Cependant un certain nombre d'échantillons indique des modifications de l'eau de mer par des sources terrestres locales. Les effets locaux sont aussi visibles au niveau des compositions isotopiques du Néodyme, qui sont en accord avec un apport provenant de roches cristallines anciennes et de sédiments du Mésozoïque, alors que d'autres échantillons indiquent une influence volcanique néogène dans le budget marin. A l'exclusion des échantillons dont les compositions correspondent à une influence locale, les compositions isotopiques du Néodyme de la Parathétys sont très similaires aux valeurs de l'Océan Indien, montrant ainsi un rôle important des masses d'eau IndoPacifiques dans cette région. Au niveau de la Méditerranée, les compositions en isotopes stables du Carbone et de l'Oxygène de foraminifères planctoniques et benthique de la région Umbria-Marche présentent un offset typique à leurs habitats. De plus, les changements dans leurs compositions suivent les changements globaux, suggérant ainsi que les conditions climatiques régionales et le cycle du carbone étaient contrôlés par des phénomènes globaux. La composition en isotopes radiogéniques d'assemblages fossiles permet une reconnaissance sur trois périodes distinctes. De 25 à 19 Ma, des rapport isotopiques élevés pour le Néodyme et faibles pour le Strontium dans les sédiments et les fossiles soutiennent l'idée d'une activité tectonique et volcanique intense, liée à l'ouverture de la Méditerranée occidentale. Entre 19 et 13 Ma, la Méditerranée présente des rapports isotopiques du Néodyme qui sont largement contrôlés par une incursion d'eau provenant de l'Océan Indien. En effet, aux alentours de 15,2 Ma, des pics dans l'enregistrement des valeurs des isotopes du Néodyme coïncident avec des événements volcaniques et de brèves diminutions du niveau marin. Finalement, de 13 à 7 Ma, les fossiles ont des rapports isotope Strontium similaires à ceux de l'eau de mer au Miocène. Les rapports isotopiques du Néodyme indiquent une influence changeante de l'océan Atlantique, et de l'océan Indien ou des sources d'eau de mer de la Parathétys qui entrent dans les bassins méditerranéens. Ce changement est guidé par des modifications globales du niveau marin et par la tectonique locale.

Constraints on Miocene oceanography and climate in the Western and Central Paratethys: O-, Sr-, and Nd-isotope compositions of marine fish and mammal remains

The Paratethys evolved as a marginal sea during the Alpine-Himalayan orogeny in the Oligo-Miocene. Sediments from the northern Alpine Molasse Basin, the Vienna, and the Pannonian Basins located in the western and central part of the Paratethys thus provide unique information on regional changes in climate and oceanography during a period of active Alpine uplift Oxygen isotope compositions of well-preserved phosphatic fossils recovered from the sediments support deposition under sub-tropical to warm-temperate climate with water temperatures of 14 to 28 degrees C for the Miocene. delta(18)O values of fossil shark teeth are similar to those reported for other Miocene marine sections and, using the best available estimates of their biostratigraphic age, show a variation until the end of the Badenian similar to that reported for composite global record. The (87)Sr/(86)Sr isotope ratios of the fossils follow the global Miocene seawater trend, albeit with a much larger scatter. The deviations of (87)Sr/(86)Sr in the samples from the well-constrained seawater curve are interpreted as due to local input of terrestrially-derived Sr. Contribution of local sources is also reflected in the epsilon(Nd) values, consistent with input from ancient crystalline rocks (e.g., Bohemian Massif and/or Mesozoic sediments with epsilon(Nd) < -9. On the other hand, there is evidence for input from areas with Neogene volcanism as suggested by samples with elevated epsilon(Nd) values >-7. Excluding samples showing local influence on the water column, an average epsilon(Nd) value of -7.9 +/- 0.5 may be inferred for the Miocene Paratethys. This value is indistinguishable from the epsilon(Nd) value of the contemporaneous Indian Ocean, supporting a dominant role of this ocean in the Western and Central Paratethys. (C) 2008 Elsevier B.V. All rights reserved.

Comparison measures of maps generated by geostatistical methods

This study uses several measures derived from the error matrix for comparing two thematic maps generated with the same sample set. The reference map was generated with all the sample elements and the map set as the model was generated without the two points detected as influential by the analysis of local influence diagnostics. The data analyzed refer to the wheat productivity in an agricultural area of 13.55 ha considering a sampling grid of 50 x 50 m comprising 50 georeferenced sample elements. The comparison measures derived from the error matrix indicated that despite some similarity on the maps, they are different. The difference between the estimated production by the reference map and the actual production was of 350 kilograms. The same difference calculated with the mode map was of 50 kilograms, indicating that the study of influential points is of fundamental importance to obtain a more reliable estimative and use of measures obtained from the error matrix is a good option to make comparisons between thematic maps.

Dependence of winter precipitation over Portugal on NAO and baroclinic wave activity

The relationship between winter (DJF) rainfall over Portugal and the variable large scale circulation is addressed. It is shown that the poles of the sea level pressure (SLP) field variability associated with rainfall variability are shifted about 15° northward with respect to those used in standard definitions of the North Atlantic Oscillation (NAO). It is suggested that the influence of NAO on rainfall dominantly arises from the associated advection of humidity from the Atlantic Ocean. Rainfall is also related to different aspects of baroclinic wave activity, the variability of the latter quantity in turn being largely dependent on the NAO. A negative NAO index (leading to increased westerly surface geostrophic winds into Portugal) is associated with an increased number of deep (ps<980 hPa) surface lows over the central North Atlantic and of intermediate (980influence on local Portuguese precipitation, but rather contribute to advection at their southern flanks. The other aspect of baroclinic wave activity varying with the NAO is the mid-tropospheric storm track (defined by the 500 hPa bandpass-filtered geopotential height variance). A possible local influence of the storm track due to vertical motions ahead of the upper air troughs cannot be unambiguously separated from the effect of advection. A separate influence of local surface cyclones over the Iberian peninsula which may, for instance, arise from the large scale ascent of air, is revealed by the statistics: for a given advection, rainfall amounts for months with local cyclone cores over the considered region tend to exceed those without. Copyright © 1999 Royal Meteorological Society

The log-exponentiated Weibull regression model for interval-censored data

In interval-censored survival data, the event of interest is not observed exactly but is only known to occur within some time interval. Such data appear very frequently. In this paper, we are concerned only with parametric forms, and so a location-scale regression model based on the exponentiated Weibull distribution is proposed for modeling interval-censored data. We show that the proposed log-exponentiated Weibull regression model for interval-censored data represents a parametric family of models that include other regression models that are broadly used in lifetime data analysis. Assuming the use of interval-censored data, we employ a frequentist analysis, a jackknife estimator, a parametric bootstrap and a Bayesian analysis for the parameters of the proposed model. We derive the appropriate matrices for assessing local influences on the parameter estimates under different perturbation schemes and present some ways to assess global influences. Furthermore, for different parameter settings, sample sizes and censoring percentages, various simulations are performed; in addition, the empirical distribution of some modified residuals are displayed and compared with the standard normal distribution. These studies suggest that the residual analysis usually performed in normal linear regression models can be straightforwardly extended to a modified deviance residual in log-exponentiated Weibull regression models for interval-censored data. (C) 2009 Elsevier B.V. All rights reserved.

Generalized log-gamma regression models with cure fraction

In this paper, the generalized log-gamma regression model is modified to allow the possibility that long-term survivors may be present in the data. This modification leads to a generalized log-gamma regression model with a cure rate, encompassing, as special cases, the log-exponential, log-Weibull and log-normal regression models with a cure rate typically used to model such data. The models attempt to simultaneously estimate the effects of explanatory variables on the timing acceleration/deceleration of a given event and the surviving fraction, that is, the proportion of the population for which the event never occurs. The normal curvatures of local influence are derived under some usual perturbation schemes and two martingale-type residuals are proposed to assess departures from the generalized log-gamma error assumption as well as to detect outlying observations. Finally, a data set from the medical area is analyzed.

Log-Burr XII regression models with censored data

In survival analysis applications, the failure rate function may frequently present a unimodal shape. In such case, the log-normal or log-logistic distributions are used. In this paper, we shall be concerned only with parametric forms, so a location-scale regression model based on the Burr XII distribution is proposed for modeling data with a unimodal failure rate function as an alternative to the log-logistic regression model. Assuming censored data, we consider a classic analysis, a Bayesian analysis and a jackknife estimator for the parameters of the proposed model. For different parameter settings, sample sizes and censoring percentages, various simulation studies are performed and compared to the performance of the log-logistic and log-Burr XII regression models. Besides, we use sensitivity analysis to detect influential or outlying observations, and residual analysis is used to check the assumptions in the model. Finally, we analyze a real data set under log-Buff XII regression models. (C) 2008 Published by Elsevier B.V.

Leverage analysis for linear mixed models

We consider a generalized leverage matrix useful for the identification of influential units and observations in linear mixed models and show how a decomposition of this matrix may be employed to identify high leverage points for both the marginal fitted values and the random effect component of the conditional fitted values. We illustrate the different uses of the two components of the decomposition with a simulated example as well as with a real data set.

Multivariate measurement error models based on scale mixtures of the skew-normal distribution

Scale mixtures of the skew-normal (SMSN) distribution is a class of asymmetric thick-tailed distributions that includes the skew-normal (SN) distribution as a special case. The main advantage of these classes of distributions is that they are easy to simulate and have a nice hierarchical representation facilitating easy implementation of the expectation-maximization algorithm for the maximum-likelihood estimation. In this paper, we assume an SMSN distribution for the unobserved value of the covariates and a symmetric scale mixtures of the normal distribution for the error term of the model. This provides a robust alternative to parameter estimation in multivariate measurement error models. Specific distributions examined include univariate and multivariate versions of the SN, skew-t, skew-slash and skew-contaminated normal distributions. The results and methods are applied to a real data set.

Diagnostic tools in beta regression with varying dispersion

We consider the issue of performing residual and local influence analyses in beta regression models with varying dispersion, which are useful for modelling random variables that assume values in the standard unit interval. In such models, both the mean and the dispersion depend upon independent variables. We derive the appropriate matrices for assessing local influence on the parameter estimates under different perturbation schemes. An application using real data is presented and discussed.