Abstract: The rural area of Peru presents a great variability of winds, its ignorance due to the absence of prediction models has an unfavorable effect on agriculture, infrastructure, security, transportation and optimization of wind energy. This work obtains an autoregressive integrated moving average (ARIMA) model for the prediction of wind speed in the R programming language for the rural area of Peru, Socabaya district of Arequipa. The methodology consists of the quantitative method and documentary technique, with a sample of 334 wind data from the year 2022 from the National Aeronautics and Space Administration (NASA) meteorological station (POWER, 2023) for the Socabaya district. Using R, the parametric methods Dickey-Fuller, Levene, D'Agostino, Kwiatkowski-Phillips-Schmidt-Shin (KPSS), and differencing (d=1) were applied to achieve normality and stationarity of the data. The simple autocorrelation function (ACF) and partial autocorrelation function (Partial ACF) are analyzed by means of a recursive adjustment process, Akaike's Information Criterion (AIC) to choose the best ARIMA prediction model. The result obtained is the ARIMA (1, 1, 2) wind prediction model, with a mean absolute scale error (MASE) precision of 0.849. It is concluded that the ARIMA model obtained can be used to predict the wind speed in Socabaya in the short term, from November 29 to December 8, 2022, and its randomness would be influenced by climate variability and amount of data from the year 2022.

Resumen: La zona rural del Perú presenta una gran variabilidad de vientos, su desconocimiento debido a la ausencia de modelos de predicción tiene un efecto desfavorable en la agricultura, infraestructura, seguridad, transporte y optimización de la energía eólica. Este trabajo obtiene un modelo de media móvil integrada autorregresiva (ARIMA) de predicción de la rapidez de viento en el lenguaje de programación R para la zona rural del Perú, distrito de Socabaya de Arequipa. La metodología consiste en el método cuantitativo y técnica documental, con una muestra de 334 datos de viento del año 2022 de la estación meteorológica Administración Nacional de Aeronáutica y Espacio (NASA) (POWER, 2023) para el distrito de Socabaya. Utilizando R, se aplicó el método paramétrico: Dicker-Fuller, Levene, D'Agostino, Kwiatkowski-Phillips-Schmidt-Shin (KPSS) y diferenciación (d=1) para lograr la normalidad y estacionariedad de los datos. Se analiza la función de autocorrelación simple (ACF) y función de autocorrelación parcial (Parcial ACF) mediante un proceso de ajuste recursivo, criterio de información de Akaike (AIC) para elegir el mejor modelo ARIMA de predicción. El resultado obtenido es el modelo de predicción de viento ARIMA (1, 1, 2), con una precisión de error de escala absoluta de media (MASE) de 0,849. Se concluye que el modelo ARIMA obtenido puede usarse para predecir la velocidad de viento en Socabaya a corto plazo, desde el 29 de noviembre hasta 8 de diciembre del 2022, y su aleatoriedad estaría influenciado por la variabilidad climática y cantidad de datos del año 2022.

Objective: to analyze the temporal pattern and estimate mortality rates in the first 24 hours of life and from preventable causes in the state of Pernambuco from 2000 to 2021. Method: an ecological study, using the quarter as the unit of analysis. The data source was made up of the Mortality Information System and the Live Birth Information System. The time series modeling was conducted according to the Autoregressive Integrated Moving Average Model. Results: 14,462 deaths were recorded in the first 24 hours of life, 11,110 (76.8%) of which being preventable. It is observed from the forecasts that the mortality rate in the first 24 hours of life ranged from 3.3 to 2.4 per 1,000 live births, and the mortality rate from preventable causes ranged from 2.3 to 1.8 per 1,000 live births. Conclusion: the prediction suggested progress in reducing mortality in the first 24 hours of life in the state and from preventable causes. The ARIMA models presented satisfactory estimates for mortality rates and preventable causes in the first 24 hours of life.

Objetivo: analisar o padrão temporal e estimar as taxas de mortalidade nas primeiras 24 horas de vida e por causas evitáveis no estado de Pernambuco no período de 2000 a 2021. Método: estudo ecológico, tendo como unidade de análise o trimestre. A fonte de dados foi constituída pelo Sistema de Informações sobre Mortalidade e pelo Sistema de Informações sobre Nascidos Vivos. A modelagem da série temporal foi conduzida segundo o Modelo Autorregressivo Integrado de Médias Móveis. Resultados: foram registrados 14.462 óbitos nas primeiras 24 horas de vida, sendo 11.110 (76,8%) evitáveis. Observa-se para os forecasts que a taxa de mortalidade nas primeiras 24 horas de vida variou de 3,3 a 2,4 por 1.000 nascidos vivos, e a taxa de mortalidade por causas evitáveis variou de 2,3 a 1,8 por 1.000 nascidos vivos. Conclusão: a previsão sugeriu avanços na redução da mortalidade nas primeiras 24 horas de vida no estado e por causas evitáveis. Os modelos ARIMA apresentaram estimativas satisfatórias para as taxas de mortalidade e por causas evitáveis nas primeiras 24 horas de vida.

Objetivo: analizar el patrón temporal y estimar las tasas de mortalidad en las primeras 24 horas de vida y por causas evitables en el estado de Pernambuco en el período de 2000 a 2021. Método: estudio ecológico, teniendo como unidad de análisis el trimestre. La fuente de datos se constituyó por el Sistema de Informaciones sobre Mortalidad y el Sistema de Informaciones sobre Nacidos Vivos. El modelado de series temporales se realizó según el Modelo Autorregresivo Integrado de Promedio Móvil. Resultados: se registraron 14.462 óbitos en las primeras 24 horas de vida, siendo 11.110 (el 76,8%) evitables. Se observa para los pronósticos ( forecasts) que la tasa de mortalidad en las primeras 24 horas de vida registro una variación de 3,3 a 2,4 por 1.000 nacidos vivos, y la tasa de mortalidad por causas evitables de 2,3 a 1,8 por 1.000 nacidos vivos. Conclusión: la predicción sugirió avances en la reducción de la mortalidad en las primeras 24 horas de vida en el estado y por causas evitables. Los modelos ARIMA presentaron estimaciones satisfactorias para las tasas de mortalidad y por causas evitables en las primeras 24 horas de vida.

Resumen La afectación en los costos y rentabilidad del negocio porcino a partir del dólar es una realidad que afecta a la economía. En este trabajo se evalúan los efectos de la Tasa Representativa del Mercado (TRM) sobre variables financieras del sector porcícola en Colombia, como los costos y el margen de utilidad del sector. Se emplea una metodología de investigación cuantitativa con enfoque correlacional, mediante pruebas de correlación de Pearson, así como modelos estadísticos, estimando los efectos de la TRM sobre el margen de utilidad y modelos de pronóstico de la TRM. Se encuentra una alta volatilidad en la TRM, además, la correlación y alta asociación con la mayoría de los costos no se puede negar, lo cual es muestra de una afectación negativa frente al margen en especial, cuando aumenta la TRM. El modelo de pronóstico de la TRM considera términos correlacionados con su propio pasado, estimando un Arima para los residuales.

Abstract This paper analyzes the impact of the exchange rate fluctuations on the financial performance of the pork sector in Colombia. The exchange rate, measured by the Representative Market Rate (RMR), affects the costs and profitability of the swine business, which is influenced by the international market. The paper uses a quantitative research methodology with a correlational approach, applying Pearson correlation tests and statistical models to estimate the effects of the RMR on the costs and profit margin of the sector, as well as to forecast the future behavior of the RMR. The paper finds that the RMR has a high volatility and a significant correlation with most of the costs, which negatively affects the profit margin, especially when the RMR increases. The paper also proposes a forecasting model for the RMR based on its own past values and an autoregressive integrated moving average (ARIMA) model for the residuals.

Abstract The Internet is a growing source of knowledge and can provide information about oral health. This ecological study aimed to characterize the interests in toothpaste among Google users from different countries. Our hypothesis was that there would be an increase in Google users’ interest in information about toothpaste. This retrospective longitudinal ecological study analyzed the toothpaste-related interest of Google users from 10 countries between January 2004 and December 2020. The monthly variation in relative search volume (RSV) and the main related queries were determined using Google Trends. Autoregressive integrated moving average (ARIMA) forecasting models were built to establish the predictive RSV values for toothpaste for an additional 12 months. Autocorrelation plots and the generalized additive model (GAM) were used to diagnose trends and seasonality in RSV curves. Additionally, the influence of social isolation related to the outbreak of COVID-19 was analyzed. Although not detected by autocorrelation function (ACF) and partial autocorrelation function (PACF) analyses, the heuristic analysis showed an increase in the interest in toothpaste-related information in all countries, with a stable trend observed in the 12-month forecasts, except for the increases in the United Kingdom and South Africa. Also, GAM analyses demonstrated a non-significant monthly or quarterly seasonal influence on data. In addition, social isolation during the COVID-19 pandemic did not influence the online information-seeking behavior of Google Search users linked to this topic. We confirmed the hypothesis that the interest of Google Search users in information about toothpaste increased in all of the 10 assessed countries.

Resumen En estadística inferencial, el pronóstico es un proceso matemático mediante el cual se hace una estimación del valor futuro de una o más variables, como puede ser la demanda. El objetivo de este presente trabajo de investigación documental fue definir la clasificación de los principales tipos de pronósticos. Además, proponer algunos de los modelos más representativos utilizados actualmente para ser implementados por pequeñas y medianas empresas, aquellos que, con base en la literatura consultada, tienen mayor potencial para lograr un pronóstico de demanda exitoso. Se encontró que los pronósticos pueden ser univariados o multivariados; sin embargo, a fin de encontrar las alternativas que impliquen menor costo para la empresa por procesamiento de datos, se consideraron exclusivamente modelos de pronósticos de series de tiempo univariados, ya que requieren únicamente los datos históricos de las ventas de la empresa. Los modelos de pronósticos de serie de tiempo se clasificaron en tres enfoques: 1) estadísticos o tradicionales, de los cuales se recomendaron el modelo de suavización exponencial triple o de Holt-Winters y el modelo de promedios móviles autorregresivos integrados (Arima), 2) de aprendizaje automático, de los cuales se destacaron el modelo bosque aleatorio y el modelo de redes neuronales recurrentes de gran memoria a corto plazo (LSTM), 3) híbridos, de los cuales se sugirieron el modelo Arima-LSTM y el modelo Facebook Prophet.

Resumo Na estatística inferencial, a previsão é um processo matemático pelo qual é feita uma estimativa do valor futuro de uma ou mais variáveis, como a demanda. O objetivo do presente trabalho de pesquisa documental foi definir a classificação dos principais tipos de previsões. Além disso, propor alguns dos modelos mais representativos atualmente utilizados para serem implementados por pequenas e médias empresas, aqueles que, com base na literatura consultada, possuem maior potencial para alcançar uma previsão de demanda bem-sucedida. Constatou-se que as previsões podem ser univariadas ou multivariadas; no entanto, para encontrar as alternativas que implicam no menor custo para a empresa para o processamento dos dados, foram considerados apenas modelos de previsão de séries temporais univariadas, uma vez que requerem apenas os dados históricos de vendas da empresa. Os modelos de previsão de séries temporais foram classificados em três abordagens: 1) estatística ou tradicional, sendo recomendado o modelo de Holt-Winters ou suavização exponencial tripla e o modelo integrado de médias móveis autorregressivas (Arima), 2) aprendizado de máquina, dos quais a floresta aleatória e o modelo de rede neural recorrente de grande memória de curto prazo (LSTM), 3) híbridos, dos quais foram sugeridos o modelo Arima-LSTM e o modelo Facebook Prophet.

Abstract In inferential statistics, forecasting is a mathematical process by which the future value of one or more variables, such as demand, is estimated. The objective of this paper was to define the classification of the main types of forecasts. In addition, to propose some of the most representative models currently used to be implemented by small and medium-sized companies, those that, based on the literature consulted, have the greatest potential to achieve a successful demand forecast. It was found that forecasts could be univariate or multivariate; however, in order to find the alternatives that would require lower cost for the enterprise by data processing, only univariate time series forecast models were recommended, as they require only historical sales data of the company. Time series forecasting models were classified into three approaches: 1) statistical or traditional, of which the Holt-Winters or triple exponential smoothing model and the autoregressive integrated moving average (Arima) model were recommended, 2) machine learning, of which the random forest model and the large short-term memory (LSTM) recurrent neural network model stood out, 3) hybrids, of which the Arima-LSTM model and the Facebook Prophet model were suggested.

Resumen El maíz es uno de los productos más importantes en el mundo debido a sus cualidades alimenticias relacionadas con el consumo humano, animal y uso industrial. Un predictor del comportamiento del precio del maíz es de utilidad para los productores y comercializadores en la toma de decisiones. Un indicador del precio es proporcionado en las bolsas internacionales; sin embargo, en México no existe una bolsa que proporcione una señal adecuada sobre el comportamiento futuro de los precios del maíz blanco en México. En esta investigación se realizó un análisis de los precios del maíz blanco en Michoacán y Jalisco, usando modelos autorregresivos integrados de media móvil (ARIMA) con el objetivo de proporcionar un predictor de los precios del maíz blanco. Se construyeron dos modelos para cada serie y se realizaron estimaciones puntuales. Se evaluó la capacidad predictiva de los modelos usando el error porcentual absoluto medio del inglés mean absolute percentage error, la raíz del error cuadrático medio y la U de Theil. Los resultados mostraron que el precio del maíz en Michoacán y Jalisco puede predecirse mediante sus valores pasados con un modelo AR (1) y un modelo MA (2). Se concluyó que estos modelos proporcionan un predictor para los precios de maíz y constituyen una herramienta útil en la planeación y toma de decisiones referentes al proceso productivo, de comercialización y productos relacionados.

Abstract Corn is one of the most important products in the world due to its nutritional qualities related to human, animal consumption and industrial use. A predictor of corn price behavior is useful for producers and marketers in making decisions. A price indicator is provided in international stock exchanges; however, in Mexico there is no stock exchange that provides an adequate signal about the future behavior of white corn prices in Mexico. In this research, an analysis of white corn prices in Michoacán and Jalisco was carried out, using autoregressive integrated moving average (ARIMA) models with the aim of providing a predictor of white corn prices. Two models were built for each series and point estimates were made. The predictive capacity of the models was evaluated using the mean absolute percentage error, the root-mean-square error and Theil’s U. The results showed that the price of corn in Michoacán and Jalisco can be predicted by its past values with an AR (1) model and an MA (2) model. It was concluded that these models provide a predictor for corn prices and constitute a useful tool in planning and making decisions regarding the process of production, commercialization and related products.

Resumo As fontes renováveis são responsáveis por mais da metade da geração elétrica brasileira, as quais correspondem basicamente às fontes hidráulica, biomassa e eólica. A presente pesquisa teve como objetivo verificar se os modelos Autorregressivos Integrados de Médias Móveis (ARIMA) possuem bom desempenho ao prever a geração de eletricidade das fontes biomassa, hidráulica e eólica nos primeiros meses da pandemia da COVID-19 no Brasil. O melhor modelo de previsão ajustado para as fontes biomassa, hidráulica e eólica foi o SARIMA, uma vez que esse modelo foi capaz de identificar os efeitos sazonais causados por instabilidades climáticas, como períodos de estiagem. Devido à sazonalidade das principais fontes geradoras, a geração renovável precisa ser compensada com outras fontes, como as não renováveis. Dessa forma, mais esforços são necessários para tornar a matriz elétrica brasileira mais sustentável.

Abstract Renewable sources are responsible for more than half of Brazilian electric generation, which basically correspond to hydropower, biomass and wind sources. This research aimed to verify if the Autoregressive Integrated Moving Average (ARIMA) models present good performance in predicting electricity generation from biomass, hydropower and wind power for the first months of COVID-19 pandemic in Brazil. The best forecasting models adjusted for biomass, hydropower and wind generation was the SARIMA, since this model was able to identify seasonal effects of climatic instability, such as periods of drought. Based on the seasonality of the largest generating sources, renewable generation needs to be offset by other sources, as non-renewable, and more efforts are needed to make Brazilian electric matrix more sustainable.

RESUMEN Introducción: Los modelos matemáticos son abstracciones de la realidad que permiten describir características importantes de los fenómenos en estudio. Entre estos los modelos epidemiológicos son considerados una de las herramientas más poderosas para analizar y comprender la propagación y el control de epidemias. Objetivos: desarrollar un modelo para el pronóstico de escenarios de contagio de la COVID-19. Métodos: Se combinaron el modelo propuesto de población inspirado en el tránsito hidrológico en ríos, un modelo autorregresivo de medias móviles (ARIMA) y un modelo de transición para evaluar tres escenarios posibles de contagio y su evolución temporal durante el desarrollo de la pandemia de la COVID-19. Resultados: Fueron estudiados 27 países con reporte de contagio hasta el 16 de abril de 2020 y fueron agrupados por cuartiles según un parámetro que relaciona personas recuperadas y fallecidas. Esta información permitió establecer criterios sobre los parámetros involucrados en el modelo propuesto para el caso de Cuba. Los escenarios de contagio de la epidemia favorable, poco favorable y crítico fueron generados a partir de modelos ARIMA hasta el 29 de abril de 2020 y de transición, extendido a 120 días. Los resultados de la implementación del modelo fueron comparados con los datos reales de la evolución de la pandemia en Cuba hasta el 22 de mayo, obteniéndose una correspondencia satisfactoria entre los registros reales y los escenarios de casos activos. Este modelo podría utilizarse de la misma forma para otros países de América que tienen también un retardo con respecto a los países seleccionados para este estudio.

ABSTRACT Introduction: The mathematical models are abstractions of the reality that allow for the description of important characteristics of the phenomena studied. Among these, the epidemic models are considered one of the most powerful tools to analyze and understand epidemic propagation and control. Objective: A model for the forecast of spread scenarios of the COVID-19 has been developed. Methods: A proposed population model inspired by analogy by hydrologic routing in rivers, an autoregressive integrated moving average model (ARIMA) and a transition model were combined in order to evaluate three possible spread scenarios and their temporary evolution during the development of COVID-19 pandemic. Results: In this contribution, 27 countries having coronavirus daily reports until April 16th, 2020 were studied. They were grouped by quartiles according to rates of recovered and deceased people. This information permitted the establishment of criteria for the parameters involved in the proposed model for Cuba as a case study. The favorable, not very favorable and critical epidemic spread scenarios were generated from ARIMA up to April 29th, 2020 and the transition model extended to 120 days. The results of the model implementation were compared with the real data of the pandemic evolution in Cuba up to May 22nd, 2020, resulting in a satisfactory correspondence between the real daily reports and the scenarios of active cases. This model could be used in the same way for other American countries that also have a delay with regard to the countries selected for this study.

Resumen: Este trabajo presenta una comparación entre dos algoritmos para la predicción de los niveles de glucosa en pacientes diabéticos mediante el uso de series de tiempo univariadas. Los algoritmos se aplican al historial de niveles de glucosa en ayunas para predecir los 5 valores posteriores. La comparación se realiza entre 1) Las Redes Neuronales Autorregresivas (ARNN) y 2) Los modelos de media móvil integrada autorregresiva (ARIMA), se analizan un total de 70 series, y se muestra que los resultados obtenidos para el modelo ARIMA tienen porcentajes de error mayores del 25% del valor predicho con respecto al valor esperado, mientras que para las Redes Neuronales Autorregresivas en el 73% de los casos el porcentaje de error fue menor al 25%.

Abstract: This work presents a comparison between two algorithms for the prediction of glucose levels in diabetic patients by using a univariate time series. The algorithms are applied to the history of fasting glucose levels to predict the five following values. The comparison is performed between 1) The Autoregressive Neural Networks (ARNN) and 2) The autoregressive integrated moving average (ARIMA) models. A total of 70 series are analyzed, and we show that the results obtained for the ARIMA model have error percentages higher than 25% of the predicted value to the expected value. In contrast, in 73% of the cases, the percentage error was less than 25% for the Autoregressive Neural Networks.

RESUMO: Objetivo: Analisar o impacto da Lei 11.705, conhecida por “Lei Seca” (LS-08), sobre a mortalidade por acidentes de trânsito (AT) em cada uma das 27 unidades federativas (UF) do Brasil. Método: Estudo ecológico de séries temporais interrompidas com dados de AT entre 2002 a 2015, totalizando 168 meses. Os dados foram obtidos do Sistema de Informações sobre Mortalidade, do Instituto Brasileiro de Geografia e Estatística e do Departamento Nacional de Trânsito. Foram ajustados modelos auto-regressivos integrados de médias móveis (ARIMA) para analisar o impacto da LS-08 em cada UF. Resultados: Após a implantação da LS-08, a mortalidade por AT diminuiu significativamente no estado de Santa Catarina (pré-LS-08 = 2,60 ± 0,30 e pós-LS-08 = 2,32 ± 0,35; p < 0,001) e no Distrito Federal (pré-LS-08 = 2,22 ± 0,40 e pós-LS-08 = 1,76 ± 0,35; p = 0,002), aumentou significativamente nos estados do Acre, Amazonas, Rondônia, Maranhão, Piauí, Ceará, Rio Grande do Norte, Paraíba, Pernambuco, Alagoas, Sergipe e Mato Grosso e permaneceu estável nos demais. Análise de sensibilidade conduzida sob uma série temporal mais curta, com 24 meses, apresentou resultados semelhantes aos obtidos com a série de 168 meses para a maioria das 27 UF. Conclusão: A LS-08 exerceu impacto heterogêneo sobre a mortalidade por AT entre as UF.

ABSTRACT: Objective: To assess the impact of 2008 Public Law number 11,705, also known as Dry Law (DL-08), on mortality from road traffic accidents (RTA), in each of the 27 Brazilian Federative Units (BFUs). Methods: Ecological study of interrupted time series with RTA data from 2002 to 2015, totalizing 168 months. Data were obtained from the Mortality Information System, the Brazilian Institute of Geography and Statistics, and from the National Traffic Department. Autoregressive integrated moving average (ARIMA) models were adjusted to assess the impact of DL-08 in each BFUs. Results: After the implementation of the DL-08, there was a significant decrease in mortality from RTA in the state of Santa Catarina (pre DL-08 = 2.60 ± 0.30 and post DL-08 = 2.32 ± 0.35; p < 0.001) and in the Federal District (pre DL-08 = 2.22 ± 0.40 and post DL-08 = 1.76 ± 0.35; p = 0.002), a significant increase in mortality in the states of Acre, Amazonas, Rondônia, Maranhão, Piauí, Ceará, Rio Grande do Norte, Paraíba, Pernambuco, Alagoas, Sergipe and Mato Grosso, and a stability in the other states. The sensitivity analysis conducted over a shorter time series with 24 months showed results similar to those obtained with the 168-month series for most of the 27 BFUs. Conclusion: The DL-08 had a heterogeneous impact on mortality from traffic accidents on BFUs.

Resumo A produção e a exportação de celulose são componentes importantes da economia no país. O objetivo desta pesquisa foi prognosticar o preço nos mercados interno e externo da celulose brasileira e avaliar a interferência entre o preço médio da celulose vendida em atacado e o preço médio da celulose exportada pelo Brasil. O estudo utilizou dados oriundos do Informativo Florestal do Centro de Estudos Avançados em Economia Aplicada (CEPEA), coletados entre junho de 2008 a março de 2018. Os modelos Autorregressivos Integrados de Médias Móveis (ARIMA) foram utilizados para prognosticar o preço da celulose em atacado e exportada pelo Brasil e, para analisar a inter-relação dessas variáveis, foi aplicado modelo Vetor Autorregressivo (VAR). Observou-se que o preço da celulose em atacado e da celulose exportada sofrem oscilações em períodos semelhantes, devido à relação direta com a cotação do dólar e com as crises financeiras nos países importadores. O modelo de melhor acurácia para prognosticar o preço da celulose em atacado foi o modelo ARIMA (1,1,0) enquanto o modelo ARFIMAX (1, d*,0) obteve o melhor desempenho para prognosticar o preço da celulose exportada. A partir da modelagem VAR, verificou-se a existência de inter-relações entre as variáveis, as quais transpareceram o forte impacto do preço da celulose em atacado sobre o preço da celulose exportada pelo Brasil. Assim, as metodologias empregadas foram eficazes para prognosticar e analisar as inter-relações entre as variáveis.

Abstract The production and export of cellulose are important components of the Brazilian economy. The aim of this research was to predict the domestic and foreign prices of the Brazilian cellulose and to evaluate the interference between its average price sold at wholesale and exported by Brazil. The study focused on data from the Forestry Newsletter of the Center for Advanced Studies on Applied Economics (CEPEA), collected from June 2008 to March 2018. The Autoregressive Integrated Moving Average (ARIMA) models were used to predict the wholesale and export price of cellulose, while the Vector Autoregressive (VAR) model was applied to analyze the inter-relation of these variables. It was observed that the price of wholesale and exported celluloses vary in similar periods, due to the direct relationship with the dollar quotation and with the financial crises in the importing countries. The most accurate model adjusted to predict the wholesale cellulose price was the ARIMA model (1,1,0), while ARFIMAX model (1, d*, 0) obtained the best performance to predict exported cellulose price. Based on the VAR model, there was a correlation between variables, which means that wholesale cellulose price has a strong impact on exported cellulose price. Thus, the methodologies used were effective to predict and analyze the inter-relationships between the variables.

Resumo Na América do Sul, a fasciolose causada pelo Trematoda Fasciola hepatica é uma antropozoonose associada a perdas econômicas significativas e baixo grau de bem-estar animal. O objetivo deste estudo foi determinar a prevalência de F. hepatica no fígado de búfalos abatidos entre 2003 a 2017 e realizar uma análise de previsão da doença para os próximos cinco anos, utilizando o modelo Auto-Regressivo Integrado de Médias Móveis (ARIMA). A análise dos dados revelou uma incidência total de 7.187 casos em 226.561 indivíduos. Houve um acentuado grau de variação interanual nas taxas de prevalência (p<0,005). Fasciola hepatica foi mais prevalente nos estados do sul do Brasil; Paraná, Rio Grande do Sul e Santa Catarina, com 11,9; 7,7; e 3,2% de fígados condenados, respectivamente. A alta incidência de condenação de fígado no Paraná foi influenciada pelo fator climático. Os modelos ARIMA indicaram uma tendência constante na ocorrência da doença, destacando um padrão futuro da doença. Os modelos também descreveram cenários de piora e de possível melhoria, calculando os efeitos de medidas de intervenção. Assim, existe a urgência de realizar diagnóstico constante nos animais (coprológico e diagnóstico imunológico) e no ambiente, para que se evite os altos índices de infecção.

Abstract In South America, fascioliasis caused by the trematode Fasciola hepatica is an anthropozoonosis disease associated with significant economic losses and poor animal welfare. The objective of this study was to determine the prevalence of F. hepatica in the liver of buffaloes slaughtered from 2003 to 2017 in Brazil, and to perform a forecast analysis of the disease for the next five years using the Autoregressive Integrated Moving Average (ARIMA) model. Data analysis revealed an incidence of 7,187 cases out of 226,561 individuals. The disease presented a considerable interannual variation (p<0.005). Fasciola hepatica was more prevalent in the southern states of Brazil; Paraná, Rio Grande do Sul, and Santa Catarina, presenting 11.9, 7.7, and 3.2% of infected livers, respectively. The high frequency of liver condemnation in Paraná was influenced by weather conditions. The ARIMA models calculated a constant trend of the disease, depicting an average of its future prevalence. The models also described a worse-case and a positive-case scenario, calculating the effects of intervention measurements. In reality, there is an urgent need for regular diagnostic in the animals (fecal and immune diagnose) and in the environment (intermediate host), in order to avoid the high rates of infection.

INTRODUCTION Dengue is a disease caused by any one of five virus serotypes and transmitted to humans by the Aedes aegypti mosquito. Climate change and health conditions have combined to make dengue a global public health problem. The situation is especially serious in Colombia, where by week 36 of 2018, dengue incidence was 96 cases per 100,000 population, with a total of 111 deaths. Different mathematical and statistical models have been proposed to understand the dynamics of transmission and consequently to apply control strategies to reduce the number of dengue cases. OBJECTIVE Forecast the number of dengue cases expected in Colombia from 2018 through 2022 with the stochastic Auto-Regressive Integrated Moving Average (ARIMA) model and use the results to adjust the parameters of an ordinary differential equations model in order to determine the disease’s basic reproduction number in the year presenting the highest number of dengue cases. METHODS An ecological time series study was conducted to forecast dengue incidence in Colombia from 2018 through 2022. The data were compiled from Colombia’s National Health Institute series on dengue cases reported by epidemiological week from 2009 to 2017. The stochastic ARIMA time series model was applied. Forecasts were then analyzed, and the year with the highest number of predicted cases was used to adjust the parameters of an ordinary differential equations model (ODE) through nonlinear least squares regression to calculate the vectorial capacity of the transmitting mosquito. RESULTS Forecasts of the total number of dengue cases per year in Colombia for the following five years were: 32,411 (2018); 88,221 (2019); 56,392 (2020); 47,940 (2021); and 77,344 (2022). The highest number of cases was forecast for 2019. Values for the parameters affecting dengue transmission that year (by the year’s four quarters), such as recovery rate (0.0992, 0.0838, 0.1177, and 0.1535, respectively), vectorial capacity of the transmitting mosquito (0.1720, 0.1705, 0.1204, and 0.2147, respectively) and the basic dengue reproduction number (1.73, 2.03, 1.02, and 1.40, respectively) were estimated, indicating that most cases would occur in the second quarter and, since the basic reproduction number values were >1, the disease would persist in the country throughout the entire year. CONCLUSIONS ARIMA model forecasts for 2018 through 2022 predicted the highest incidence of dengue cases in Colombia would occur in 2019. Comparison of ARIMA model forecasts and the ODE model allowed projections of possible variations in dengue cases reported, and the basic reproduction number predicted that dengue would persist throughout 2019.

This paper proposes and discusses Autoregressive Moving Average (ARMA), Autoregressive Integrated Moving Average (ARIMA) and Seasonal Autoregressive Integrated Moving Average (SARIMA) time series models for broadband power line communication (PLC) networks with impulsive noise enviroment in the frequency range of 1 - 30 MHz. In time series modelling and analysis, time series models are fitted to the acquired time series describing the system for purposes which include simulation, forecasting, trend assessment, and a better understanding of the dynamics of the impulsive noise in PLC systems. Also, because the acquired impulsive noise measurement data are observations made over time, time series models constitute important statistical tools for use in solving the problem of impulsive noise modelling and forecasting in PLC. In fact, the time series and other statistical methods presented in numerous available literature draw upon research developments from two areas of environmetrics called stochastic hydrology and statistical water quality modelling as well as research contributions from the field of statistics. In time series modelling and analysis, we determine the most appropriate stochastic or time series model to fit our acquired data set at the confirmatory data analysis stage. No matter what type of stochastic model is to be fitted to the data set, we follow the identification, estimation, and diagnostic check stages of model construction. In addition, we explore the resulting autocorrelation functions in estimating the parameters of the selected time series models. Finally, SARIMA model is found suitable for computer-based PLC systems simulations and forecasting based on the diagnostic checks.

Resumo Esse trabalho tem como objetivo realizar previsões de séries temporais da velocidade do vento em termos de médias mensais no nordeste brasileiro. Foram testados os seguintes modelos de séries temporais, Auto-Regressivo Integrado de Médias Móveis (ARIMA) e Holt–Winters (HW), e também inteligência artificial computacional com o uso de Redes Neurais Artificiais (RNA). E ainda foram testados dois modelos híbridos, o primeiro com combinação dos modelos ARIMA e RNA, o qual existem trabalhos na literatura, e o segundo trata-se de uma tentativa de combinação dos modelos HW e RNA. As séries ajustadas obtidas pelos modelos híbridos, mostram-se eficientes em acompanhar o perfil das séries observadas das regiões de estudo, com semelhanças aos dados observados em termos de máximos e mínimos, indicando assim, a capacidade dos modelos em representar sazonalidades. O cálculo da estatística de erros envolvendo o modelo híbrido (HW e RNA) obteve os menores valores em Fortaleza, São Luís e Natal, por exemplo, com valores de erros percentuais de 3,80%, 4,91% e 2,85%, respectivamente. A redução das variáveis da estatística de erros por parte dos modelos híbridos quando comparado separadamente a utilização dos modelos (ARIMA, HW e RNA) poderá influenciar nas precisões das velocidades dos ventos previstas.

Abstract The objective of this work is to perform time series forecasts of wind speed in terms of monthly averages in the Brazilian Northeast. The following time-series models, Auto-Regressive Integrated Moving Average (ARIMA) and Holt-Winters (HW) were tested, as well as computational artificial intelligence with the use of Artificial Neural Networks (ANN). In addition, two hybrid models were tested, the first with a combination of the ARIMA and ANN models, which work in the literature, and the second is an attempt to combine HW and ANN models. The adjusted series obtained by the hybrid models, are efficient to follow the profile of the observed series of the study regions, with similarities to the data observed in terms of maxima and minima, thus indicating the capacity of the models to represent seasonalities. The calculation of error statistics involving the hybrid (HW and ANN) model obtained the lowest values in Fortaleza, São Luís and Natal, for example, with percent error values of 3.80%, 4.91% and 2.85%, respectively. The reduction of the statistical variables of errors by the hybrid models when comparing the use of the models (ARIMA, HW and ANN) separately may influence the predictions of expected wind velocities.