Autoregressive models have played an important role in time series. In this paper, an autoregressive model based on the skew-normal distribution is considered. The estimation of its parameters is carried out by using the expectation–maximization algorithm, whereas the diagnostic analytics are conducted by means of the local influence method. Normal curvatures for the model under four perturbation schemes are established. Simulation studies are conducted to evaluate the performance of the proposed procedure. In addition, an empirical example involving weekly financial return data are analyzed using the procedure with the proposed diagnostic analytics, which has improved the model fit.
Introduction Suicide is the second leading cause of premature death in people between 15 and 29 years old and the third in young people between 15 and 19 years old. Adolescence is a critical period concerning mental health disorders since there is greater vulnerability to suicidal behaviors. The situation in Latin America is worrying, with Chile being one of the two countries where suicide rates of children and adolescents increase yearly. This study aims to analyze clinical, psychological, family, and social risk factors associated with suicidal behavior in a clinical sample of adolescents treated in the public health system of the Maule region. Methods The study design is cross- sectional. We used a sample of 388 adolescents between 10 and 21 years old admitted to the health system of the Maule Region. The participants were evaluated by applying five measuring instruments (The Barrat Impulsivity Scale, The Difficulties in Emotional Regulation Scale, The Depression, Anxiety and Stress Scale, The General Help-Seeking Questionnaire for mental health problems in adolescents, and The Columbia Suicide Severity Rating Scale) in addition to collecting social and family information and relevant clinical history from the medical records. Results The analysis allowed us to identify distinctive characteristics of adolescent suicidal behavior by describing clinical, psychological, and family social factors. Conclusions Adolescents with a history of suicide attempts are characterized by having suicidal ideation, anxious-depressive symptoms, stress, insomnia, and impulsiveness. Likewise, they report being non-religious, belonging to sexual minorities, and victims of sexual harassment and/or abuse.
Longitudinal count responses are often analyzed with a Poisson mixed model. However, under overdispersion, these responses are better described by a negative binomial mixed model. Estimators of the corresponding parameters are usually obtained by the maximum likelihood method. To investigate the stability of these maximum likelihood estimators, we propose a methodology of sensitivity analysis using local influence. As count responses are discrete, we are unable to perturb them with the standard scheme used in local influence. Then, we consider an appropriate perturbation for the means of these responses. The proposed methodology is useful in different applications, but particularly when medical data are analyzed, because the removal of influential cases can change the statistical results and then the medical decision. We study the performance of the methodology by using Monte Carlo simulation and applied it to real medical data related to epilepsy and headache. All of these numerical studies show the good performance and potential of the proposed methodology.
Autoregressive models in time series are useful in various areas. In this article, we propose a skew-t autoregressive model. We estimate its parameters using the expectation-maximization (EM) method and develop the influence methodology based on local perturbations for its validation. We obtain the normal curvatures for four perturbation strategies to identify influential observations, and then to assess their performance through Monte Carlo simulations. An example of financial data analysis is presented to study daily log-returns for Brent crude futures and investigate possible impact by the COVID-19 pandemic.
17 June 2020 in Journal of Statistical Computation and Simulation
This article proposes a methodology for diagnostics in a logistic regression with random intercept motivated by a biological study. The methodology includes local and global influence techniques allowing us to contrast the results of both types of influence. The proposed methodology is applied to a case study with real data to show its potential. This study corresponds to the reproduction of arachnids reporting how the local and global influence of atypical observations can modify the significance of parameters, and then the biological conclusions. The model fitting is evaluated through predictive indicators. The methodology is summarized in an algorithm and a demo example is implemented in R code to facilitate its application. To evaluate the performance of the methodology, Monte Carlo simulations are conducted.