Transformed generalized linear models

The estimation of data transformation is very useful to yield response variables satisfying closely a normal linear model, Generalized linear models enable the fitting of models to a wide range of data types. These models are based on exponential dispersion models. We propose a new class of transformed generalized linear models to extend the Box and Cox models and the generalized linear models. We use the generalized linear model framework to fit these models and discuss maximum likelihood estimation and inference. We give a simple formula to estimate the parameter that index the transformation of the response variable for a subclass of models. We also give a simple formula to estimate the rth moment of the original dependent variable. We explore the possibility of using these models to time series data to extend the generalized autoregressive moving average models discussed by Benjamin er al. [Generalized autoregressive moving average models. J. Amer. Statist. Assoc. 98, 214-223]. The usefulness of these models is illustrated in a Simulation study and in applications to three real data sets. (C) 2009 Elsevier B.V. All rights reserved.

Transformed symmetric models

For the first time, we introduce a class of transformed symmetric models to extend the Box and Cox models to more general symmetric models. The new class of models includes all symmetric continuous distributions with a possible non-linear structure for the mean and enables the fitting of a wide range of models to several data types. The proposed methods offer more flexible alternatives to Box-Cox or other existing procedures. We derive a very simple iterative process for fitting these models by maximum likelihood, whereas a direct unconditional maximization would be more difficult. We give simple formulae to estimate the parameter that indexes the transformation of the response variable and the moments of the original dependent variable which generalize previous published results. We discuss inference on the model parameters. The usefulness of the new class of models is illustrated in one application to a real dataset.

MODELLING AND FORECASTING INFLATION RATES IN GHANA: AN APPLICATION OF SARIMA MODELS

Ghana faces a macroeconomic problem of inflation for a long period of time. The problem in somehow slows the economic growth in this country. As we all know, inflation is one of the major economic challenges facing most countries in the world especially those in African including Ghana. Therefore, forecasting inflation rates in Ghana becomes very important for its government to design economic strategies or effective monetary policies to combat any unexpected high inflation in this country. This paper studies seasonal autoregressive integrated moving average model to forecast inflation rates in Ghana. Using monthly inflation data from July 1991 to December 2009, we find that ARIMA (1,1,1)(0,0,1)12 can represent the data behavior of inflation rate in Ghana well. Based on the selected model, we forecast seven (7) months inflation rates of Ghana outside the sample period (i.e. from January 2010 to July 2010). The observed inflation rate from January to April which was published by Ghana Statistical Service Department fall within the 95% confidence interval obtained from the designed model. The forecasted results show a decreasing pattern and a turning point of Ghana inflation in the month of July.

A family of autoregressive conditional duration models

This paper develops a family of autoregressive conditional duration (ACD) models that encompasses most specifications in the literature. The nesting relies on a Box-Cox transformation with shape parameter λ to the conditional duration process and a possibly asymmetric shocks impact curve. We establish conditions for the existence of higher-order moments, strict stationarity, geometric ergodicity and β-mixing property with exponential decay. We next derive moment recursion relations and the autocovariance function of the power λ of the duration process. Finally, we assess the practical usefulness of our family of ACD models using NYSE transactions data, with special attention to IBM price durations. The results warrant the extra flexibility provided either by the Box-Cox transformation or by the asymmetric response to shocks.

A family of autoregressive conditional duration models

This paper develops a family of autoregressive conditional duration (ACD) models that encompasses most specifications in the literature. The nesting relies on a Box-Cox transformation with shape parameter λ to the conditional duration process and a possibly asymmetric shocks impact curve. We establish conditions for the existence of higher-order moments, strict stationarity, geometric ergodicity and β-mixing property with exponential decay. We next derive moment recursion relations and the autocovariance function of the power λ of the duration process. Finally, we assess the practical usefulness of our family of ACD models using NYSE price duration data on the IBM stock. The results warrant the extra flexibility provided either by the Box-Cox transformation or by the asymmetric response to shocks.

Effects of the essential oil from Citrus aurantium L. in experimental anxiety models in mice

Citrus aurantium L. is popularly used to treat anxiety, among other indications suggesting central nervous system action. Previous studies showed anxiolytic effect in the essential oil from peel in mice evaluated on the elevated plus maze [Carvalho-Freitas, M.I.R., Costa, M., 2002. Anxiolytic and sedative effects of extracts and essential oil from Citrus aurantium L. Biological and Pharmaceutical Bulletin 25, 1629-1633.]. In order to better characterize the activity of the essential oil, it was evaluated in two other experimental models: the light-dark box and the marble-burying test, respectively related to generalized anxiety disorder and to obsessive compulsive disorder. Mice were treated acutely by oral route 30 min (single dose) or once a day for 15 days (repeated doses) before experimental procedures. In light-dark box test, single treatment with essential oil augmented the time spent by mice in the light chamber and the number of transitions between the two compartments. There were no observed alterations in the parameters evaluated in light-dark box after repeated treatment. Otherwise, single and repeated treatments with essential oil were able to suppress marble-burying behavior. At effective doses in the behavioral tests, mice showed no impairment on rotarod procedure after both single and repeated treatments with essential oil, denoting absence of motor deficit. Results observed in marble-burying test, related to obsessive compulsive disorder, appear more consistent than those observed in light-dark box. (c) 2005 Elsevier B.V. All rights reserved.

Fractional polynomial response surface models

Second-order polynomial models have been used extensively to approximate the relationship between a response variable and several continuous factors. However, sometimes polynomial models do not adequately describe the important features of the response surface. This article describes the use of fractional polynomial models. It is shown how the models can be fitted, an appropriate model selected, and inference conducted. Polynomial and fractional polynomial models are fitted to two published datasets, illustrating that sometimes the fractional polynomial can give as good a fit to the data and much more plausible behavior between the design points than the polynomial model. © 2005 American Statistical Association and the International Biometric Society.

Previsão estocástica de nível fluviométrico para cidade de Marabá-PA: método de Box-Jenkins

O Município de Marabá- PA, situado na região Amazônica, sudeste do Estado do Pará, sofre anualmente com eventos de enchentes, ocasionados pelo aumento periódico do rio Tocantins e pela situação de vulnerabilidade da população que reside em áreas de risco. A defesa civil estadual e municipal anualmente planeja e prepara equipes para ações de defesa no município. Nesta fase o monitoramento e previsão de eventos de enchentes são importantes. Portanto, com o objetivo de diminuir erros nas previsões hidrológicas para o Município de Marabá, desenvolveu-se um modelo estocástico para previsão de nível do rio Tocantins, baseado na metodologia de Box e Jenkins. Utilizou os dados de níveis diários observados nas estações hidrológicas de Marabá e Carolina e Conceição do Araguaia da Agência Nacional de Águas (ANA), do período de 01/12/ 2008 a 31/03/2011. Efetuou-se o ajustamento de três modelos (M_t, N_t e Y_t), através de diferentes aplicativos estatísticos: o SAS e o Gretl, usando diferentes interpretações do comportamento das séries para gerar as equações dos modelos. A principal diferença entre os aplicativos é que no SAS usa o modelo de função de transferência na modelagem. Realizou-se uma classificação da variabilidade do nível do rio, através da técnica dos Quantis para o período de 1972 a 2011, examinando-se apenas as categorizações de níveis ACIMA e MUITO ACIMA do normal. Para análise de impactos socioeconômicos foram usados os dados das ações da Defesa Civil Estado do Pará nas cheias de 2009 e 2011. Os resultados mostraram que o número de eventos de cheias com níveis MUITO ACIMA do normal, geralmente, podem estar associados a eventos de La Niña. Outro resultado importante: os modelos gerados simularam muito bem o nível do rio para o período de sete dias (01/04/2011 a 07/04/2011). O modelo multivariado N_t (com pequenos erros) representou o comportamento da série original, subestimando os valores reais nos dias 3, 4 e 5 de abril de 2011, com erro máximo de 0,28 no dia 4. O modelo univariado (Y_t) teve bons resultados nas simulações com erros absolutos em torno de 0,12 m. O modelo com menor erro absoluto (0,08m) para o mesmo período foi o modelo M_t, desenvolvido pelo aplicativo SAS, que interpreta a série original como sendo não linear e não estacionária. A análise quantitativa dos impactos fluviométricos, ocorridos nas enchentes de 2009 e 2011 na cidade de Marabá, revelou em média que mais de 4 mil famílias sofrem com estes eventos, implicado em gastos financeiros elevados. Logo, conclui-se que os modelos de previsão de níveis são importantes ferramentas que a Defesa Civil, utiliza no planejamento e preparo de ações preventivas para o município de Marabá.

Previsão da velocidade dos ventos por redes neurais artificiais e arima de box & jenkins

Pós-graduação em Engenharia Elétrica - FEIS

The bias in reversing the Box-Cox transformation in time series forecasting: An empirical study based on neural networks

The Box-Cox transformation is a technique mostly utilized to turn the probabilistic distribution of a time series data into approximately normal. And this helps statistical and neural models to perform more accurate forecastings. However, it introduces a bias when the reversion of the transformation is conducted with the predicted data. The statistical methods to perform a bias-free reversion require, necessarily, the assumption of Gaussianity of the transformed data distribution, which is a rare event in real-world time series. So, the aim of this study was to provide an effective method of removing the bias when the reversion of the Box-Cox transformation is executed. Thus, the developed method is based on a focused time lagged feedforward neural network, which does not require any assumption about the transformed data distribution. Therefore, to evaluate the performance of the proposed method, numerical simulations were conducted and the Mean Absolute Percentage Error, the Theil Inequality Index and the Signal-to-Noise ratio of 20-step-ahead forecasts of 40 time series were compared, and the results obtained indicate that the proposed reversion method is valid and justifies new studies. (C) 2014 Elsevier B.V. All rights reserved.

Lipopolysaccharide-Induced Sickness Behaviour Evaluated in Different Models of Anxiety and Innate Fear in Rats

The fact that there is a complex and bidirectional communication between the immune and nervous systems has been well demonstrated. Lipopolysaccharide (LPS), a component of gram-negative bacteria, is widely used to systematically stimulate the immune system and generate profound physiological and behavioural changes, also known as sickness behaviour (e.g. anhedonia, lethargy, loss of appetite, anxiety, sleepiness). Different ethological tools have been used to analyse the behavioural modifications induced by LPS; however, many researchers analysed only individual tests, a single LPS dose or a unique ethological parameter, thus leading to disagreements regarding the data. In the present study, we investigated the effects of different doses of LPS (10, 50, 200 and 500 mu g/kg, i.p.) in young male Wistar rats (weighing 180200 g; 89 weeks old) on the ethological and spatiotemporal parameters of the elevated plus maze, light-dark box, elevated T maze, open-field tests and emission of ultrasound vocalizations. There was a dose-dependent increase in anxiety-like behaviours caused by LPS, forming an inverted U curve peaked at LPS 200 mu g/kg dose. However, these anxiety-like behaviours were detected only by complementary ethological analysis (stretching, grooming, immobility responses and alarm calls), and these reactions seem to be a very sensitive tool in assessing the first signs of sickness behaviour. In summary, the present work clearly showed that there are resting and alertness reactions induced by opposite neuroimmune mechanisms (neuroimmune bias) that could lead to anxiety behaviours, suggesting that misunderstanding data could occur when only few ethological variables or single doses of LPS are analysed. Finally, it is hypothesized that this bias is an evolutionary tool that increases animals security while the body recovers from a systemic infection.

Paleoenvironment and Paleoecology of a Late Paleocene High-Latitude Terrestrial Succession, Arkose Ridge Formation at Box Canyon, Southern Talkeetna Mountains, Alaska

Paleogene sedimentary rocks of the Arkose Ridge Formation (Talkeetna Mountains, Alaska) preserve a record of a fluvial-lacustrine depositional environment and its forested ecosystem in an active basin among the convergent margin tectonic processes that shaped southern Alaska. An -800 m measured succession at Box Canyon indicates braid-plain deposition with predominantly gravelly deposits low in the exposure to sandy and muddy facies associations below an overlying lava flow sequence. U-Pb geochronology on zircons from a tuff and a sandstone within the measured section, as well as an Ar/Ar date from the overlying lava constrain the age of the sedimentary succession to between similar to 59 Ma and 48 Ma Fossil plant remains occur throughout the Arkose Ridge Formation as poorly-preserved coalified woody debris and fragmentary leaf impressions. At Box Canyon, however, a thin la-custrine depositional lens of rhythmically laminated mudrocks yielded fish fossils and a well-preserved floral assemblage including foliage and reproductive organs representing conifers, sphenopsids, monocots, and dicots. Leaf physiognomic methods to estimate paleoclimate were applied to the dicot leaf collection and indicate warm temperate paleotemperatures (-11-15 +/- -4 degrees C MAT) and elevated paleoprecipitation (-120 cm/yr MAP) estimates as compared to modem conditions; results that are parallel with previously published estimates from the partly coeval Chickaloon Formation deposited in more distal depositional environments in the same basin. The low abundance of leaf herbivory in the Box Canyon dicot assemblage (-9% of leaves damaged) is also similar to the results from assemblages in the meander-plain depositional systems of the Chickaloon. This new suite of data informs models of the tectonostratigraphic evolution of southern Alaska and the developing understanding of terrestrial paleoecology and paleoclimate at high latitudes during the Late Paleocene-Early Eocene greenhouse climate phase. (c) 2014 Elsevier B.V. All rights reserved.

Age models and stable isotope analysis on sediment core Site 199-1218 from the equatorial Pacific

A 13-million-year continuous record of Oligocene climate from the equatorial Pacific reveals a pronounced "heartbeat" in the global carbon cycle and periodicity of glaciations. This heartbeat consists of 405,000-, 127,000-, and 96,000-year eccentricity cycles and 1.2-million-year obliquity cycles in periodically recurring glacial and carbon cycle events. That climate system response to intricate orbital variations suggests a fundamental interaction of the carbon cycle, solar forcing, and glacial events. Box modeling shows that the interaction of the carbon cycle and solar forcing modulates deep ocean acidity as well as the production and burial of global biomass. The pronounced 405,000-year eccentricity cycle is amplified by the long residence time of carbon in the oceans.

Age models and sea-surface paleotemperature reconstruction of sediment cores from the eastern equatorial Atlantic

Based on the faunal record of planktonic foraminifers in three long gravity sediment cores from the eastern equatorial Atlantic, the sea-surface temperature history ove the last 750,000 years was studied at a resolution of 3,000 to 10,000 years. Detailed oxygen-isotope and paleomagnetic stratigraphy helped to identify the following major faunal events: Globorotaloides hexagonus and Globorotalia tumida flexuosa became extinct in the eastern tropical Atlantic at the isotope stage 4/5 boundary, now dated at 68,000 years B.P. The persistent occurrence of the pink variety of Globigerinoides ruber started during the late stage 12 at 410,000 years B.P. CARTUNE-age. This datum may provide an easily detectible faunal stratigraphic marker for the mid-Brunhes Chron. The updated scheme of the Ericson zones helped the recognition of a hiatus at the northwestern slope of the Sierra Leone Basin covering oxygen-isotope stages 10 to 12. Classifying the planktonic foraminifer counts into six faunal assemblages, according to the factor analysis derived model of Pflaumann (1985), the tropical and the tropical-upwelling communities account for 57 % at Site 16415, and 86 % at Site 13519, respectively of the variance of the faunal record. A largely continuous paleotemperature record for both winter and summer seasons was obtained from the top of the Sierra Leone Rise with the winter temperatures ranging between 20 and 25 °C, and the summer ones between 24 and 30 °C. The record of cores from greater water depths is frequently interrupted by samples with no-analogue faunal communities and/or poor preservation. Based on the seasonality signal, during cold periods the termal equator shifted to a geographically mnore asymmetrical northern position. Dissolution altering the faunal communities becomes stronger with greater water depth, the estimated mean minimum loss of specimens increases from 70 % to 80 % between 2,860 and 3,850 water depth although some species will be more susceptible than others. Enhanced dissolution occured during stage 4 but also during cold phases in the warm stage 7 and 9. Correlations between the Foraminiferal Dissolution Index and the estimated sea-surface temperatures are significant. Foraminiferal flux rates, negatively correlated to the flux rates of organic carbon and of diatoms, may be a result of enhanced dissolution during cold stages, destroying still more of the faunal signal than indicated by the calculated minimum loss. The fluctuations of the oxygen-isotope curves and the hibernal sea-surfave temperatures are fairly coherent. During warm oxygen-isotope stages the temperature maxima lag often by 5 to 15 ka behind the respective sotope minima. During cold stages, sea-surface temperature changes are partly out of phase and contain additional fluctuations.

Reduced order adaptive models for systems of PDEs using POD

In this work we propose a method to accelerate time dependent numerical solvers of systems of PDEs that require a high cost in computational time and memory. The method is based on the combined use of such numerical solver with a proper orthogonal decomposition, from which we identify modes, a Galerkin projection (that provides a reduced system of equations) and the integration of the reduced system, studying the evolution of the modal amplitudes. We integrate the reduced model until our a priori error estimator indicates that our approximation in not accurate. At this point we use again our original numerical code in a short time interval to adapt the POD manifold and continue then with the integration of the reduced model. Application will be made to two model problems: the Ginzburg-Landau equation in transient chaos conditions and the two-dimensional pulsating cavity problem, which describes the motion of liquid in a box whose upper wall is moving back and forth in a quasi-periodic fashion. Finally, we will discuss a way of improving the performance of the method using experimental data or information from numerical simulations