2

Over the last decades, the challenges in survival models have been changing consider- ably and full probabilistic modeling is crucial in many medical applications. Motivated from a new biological interpretation of cancer metastasis, we introduce a general method for obtaining more flexible cure rate models. The proposal model extended the promotion time cure rate model. Furthermore, it includes several well-known mod- els as special cases and defines many new special models. We derive several properties of the hazard function for the proposed model and establish mathematical relation- ships with the promotion time cure rate model. We consider a frequentist approach to perform inferences, and the maximum likelihood method is employed to estimate the model parameters. Simulation studies are conducted to evaluate its performance with a discussion of the obtained results. A real dataset from population-based study of incident cases of melanoma diagnosed in the state of São Paulo, Brazil, is discussed in detail.

In this article, we introduce a long-term survival model in which the number of competing causes of the event of interest follows the zero-modified geometric (ZMG) distribution. Such distribution accommodates equidispersion, underdispersion, and overdispersion and captures deflation or inflation of zeros in the number of lesions or initiated cells after the treatment. The ZMG distribution is also an appropriate alternative for modeling clustered samples when the number of competing causes of the event of interest consists of two subpopulations, one containing only zeros (cure proportion), while in the other (noncure proportion) the number of competing causes of the event of interest follows a geometric distribution. The advantage of this assumption is that we can measure the cure proportion in the initiated cells. Furthermore, the proposed model can yield greater or lower cure proportion than that of the geometric distribution when modeling the number of competing causes. In this article, we present some statistical properties of the proposed model and use the maximum likelihood method to estimate the model parameters. We also conduct a Monte Carlo simulation study to evaluate the performance of the estimators. We present and discuss two applications using real-world medical data to assess the practical usefulness of the proposed model.

This work tests the null hypothesis that the coefficients of the total landings, landed values, mean catches and price per kg of migratory and resident species are constant over time following the installation of two large run- of-the-river hydroelectric dams in a large tropical river. To identify shifts in catches and economic returns due to river impoundment, we inspected daily landing data (25-year time series) and wholesale prices (19-year time series) for the Madeira River, the largest tributary of the Amazon River. Our results show that the period of decreasing catches and increasing prices observed for fisheries in the Madeira River matched the timings of the construction of the two dams. According to the results, both dams quickly changed catches and fish supply to market, which were immediately echoed in the price per kg of exploited fish species. Following the dam con- struction, prices rose for both fish that became scarce and fish that became abundant. Though catches declined 58% in 25 years, the price increased 49% over the same period, representing a high economic cost for the local population. Further, there was a clear decline in the catches of some species (e.g., the dourada and the curimatã), but increased catches of others (e.g., the sardine and the tucunaré). Moreover, some fluctuation patterns across years showed natural oscillations, or changes, in local habitats and even fishing efforts.

Neste artigo, argumentamos que modelos lineares mistos (MLMs) são mais adequados que Análise de Variância (ANOVA) para lidar com dados de tempo de reação. Com a análise de um conjunto de dados simulados, mostramos que MLMs têm menor chance de incorrer em erros do Tipo I por permitir a inclusão de mais de um efeito aleatório (geralmente participantes e itens) em um único modelo. Também apresentamos uma introdução à implementação e análise de dados por meio de MLMs usando R e sugerimos materiais adicionais para os pesquisadores que desejarem fazer esse tipo de análise. Nosso principal objetivo é fomentar o uso de MLMs na comunidade psicolinguística brasileira.

Maximum analysis consists of modeling the maximums of a data set by con- sidering a specific distribution. Extreme value theory (EVT) shows that, for a sufficiently large block size, the maxima distribution is approximated by the gen- eralized extreme value (GEV) distribution. Under EVT, it is important to observe the high quantiles of the distribution. In this sense, quantile regression tech- niques fit the data analysis of maxima by using the GEV distribution. In this context, this work presents the quantile regression extension for the GEV distri- bution. In addition, a time-varying quantile regression model is presented, and the important properties of this approach are displayed. The parameter estima- tion of these new models is carried out under the Bayesian paradigm. The results of the temperature data and river quota application show the advantage of using this model, which allows us to estimate directly the quantiles as a function of the covariates. This shows which of them influences the occurrence of extreme temperature and the magnitude of this influence.

We present a novel first-order nonnegative integer-valued autoregressive model for stationary count data processes with Bernoulli-geometric marginals based on a new type of generalized thinning operator. It can be used for modeling time series of counts with equidispersion, underdispersion and overdispersion. The main properties of the model are derived, such as probability generating function, moments, transition probabilities and zero probability. The maximum likelihood method is used for estimating the model parameters. The proposed model is fitted to time series of counts of iceberg orders and of cases of family violence illustrating its capabilities in challenging cases of overdispersed and equidispersed count data.

We introduce a general method for obtaining more flexible new distributions by com- pounding the extended Weibull and power series distributions. The compounding proce- dure follows the same set-up carried out by Adamidis and Loukas (1998) and defines 68 new sub-models. The new class of generated distributions includes some well-known mix- ing distributions, such as the Weibull power series (Morais and Barreto-Souza, 2011) and exponential power series (Chahkandi and Ganjali, 2009) distributions. Some mathematical properties of the new class are studied including moments and the generating function. We provide the density function of the order statistics and their moments. The method of maximum likelihood is used for estimating the model parameters. Special distributions are investigated. We illustrate the usefulness of the new distributions by means of two appli- cations to real data sets.

Our goal in this work is to construct empirical confidence intervals for the fractional parameter $d$ in ARFIMA $(0, d, 0)$ processes. Through these confidence intervals one can compare several estimators for d to decide which one is the best estimation method related to long memory time series. We use a fortran routine that simulates random time series to later perform an analysis for detecting long memory. We also apply the methodology to real DNA sequences to evaluate the efficiency of our method in the construction of these confidence intervals.