Journal of Statistical Sciences

Fuzzy Lee-Carter Model in Mortality Data Analysis

Jalal Chachi

The Lee-Carter model is a useful dynamic stochastic model representing the evolution of central mortality rates over time. This model only considers the uncertainty about the coefficient related to the mortality trend over time but not the age-dependent coefficients. This paper proposes a fuzzy extension of the Lee-Carter model that allows quantifying the uncertainty of both kinds of parameters. The variability of the time-dependent index is modeled as a stochastic fuzzy time series. Likewise, the uncertainty of the age-dependent coefficients is quantified using triangular fuzzy numbers. Considering this last hypothesis requires developing and solving a fuzzy regression model. Once the generalization of the desired fuzzy model is introduced, we will show how to fit the logarithm of the central mortality rate in Khuzestan province using by using fuzzy numbers arithmetic during the years 1401-1383 and random fuzzy forecast in the years 1402-1406.

Comparison of Neyman-Pearson and Evidential Approaches in Separate One-Sided Tests Based on Model Misspecification Error

Ali Dastbaravarde

In statistical hypothesis testing, model misspecification error occurs when the real model of the data is none of the models under null and alternative hypotheses. This research has studied the probability of model misspecification errors in one-sided tests. These error rates are compared between the Neyman-Pearson and evidential statistical inference approaches. The results show that the evidential approach works better than the Neyman-Pearson approach.

Early Detection of Down Syndrome Using Machine Learning Algorithms

Abdolreza Sayyareh

Non-invasive NIPT test has been used in trisomy 21 screening. However, there is a possibility of misdiagnosis in the methods used to diagnose Down syndrome. Therefore, it is essential to provide a process that can be used alongside these methods to improve efficiency. The main goal of this article is to design a model based on machine learning algorithms for the early diagnosis of Down syndrome. Machine learning algorithms such as support vector machine, simple Bayes, decision tree, random forest, and nearest neighbor, which are frequently used to improve the diagnosis of disorders, have been implemented on the mentioned dataset. The performance of each model on the Down syndrome dataset was investigated, and the most suitable model for this purpose was introduced.

Autoregressive Spatial Regression Model and Second-Order Moving Average for Generalized Skew-Laplace Random Field

Mohammad Mehdi Saber

In this article, autoregressive spatial regression and second-order moving average will be presented to model the outputs of a heavy-tailed skewed spatial random field resulting from the developed multivariate generalized Skew-Laplace distribution. The model parameters are estimated by the maximum likelihood method using the Kolbeck-Leibler divergence criterion. Also, the best spatial predictor will be provided. Then, a simulation study is conducted to validate and evaluate the performance of the proposed model. The method is applied to analyze a real data.

Prediction for Coherent System Lifetime Based on Type-II Censored Data from Half Logistic Distribution

roshanak zaman

In this paper, the prediction of the lifetime of k-component coherent systems is studied using classical and Bayesian approaches with type-II censored system lifetime data. The system structure and signature are assumed to be known, and the component lifetime distribution follows a half-logistic model. Various point predictors, including the maximum likelihood predictor, the best-unbiased predictor, the conditional median predictor, and the Bayesian point predictor under a squared error loss function, are calculated for the coherent system lifetime. Since the integrals required for Bayes prediction do not have closed forms, the Metropolis-Hastings algorithm and importance sampling methods are applied to approximate these integrals. For type-II censored lifetime data, prediction interval based on the pivotal quantity, prediction interval HCD, and Bayesian prediction interval are considered. A Monte Carlo simulation study and a numerical example are conducted to evaluate and compare the performances of the different prediction methods.

Equivalents of Tail Risk Measures Based on Inequality Indexes in the Economy and Reliability

Gholam reza Mohtashami Borzadaran

Abstract: The use of tail risk measures has been noticed in recent decades, especially in the financial and banking industry. The most common ones are value at risk and expected shortfall. The tail Gini risk measure, a composite risk measure, was introduced recently. The primary purpose of this article is to find the relationship between the concepts of economic risks, especially the expected shortfall and the tail Gini risk measure, with the concepts of inequality indices in the economy and reliability. Examining the relationship between these concepts allows the researcher to use the concepts of one to investigate other concepts. As you will see below, the existing mathematical relationships between the tail risk measures and the mentioned indices have been obtained, and these relationships have been calculated for some distributions. Finally, real data from the Iranian Stock Exchange was used to familiarize the concept of this tail risk measure.

Variational Bayesian Analysis of Skew Spatial Regression Model Based on a flexible Subclass of Closed Skew-Normal Distribution

Omid Karimi

Spatial regression models are used to analyze quantitative spatial responses based on linear and non-linear relationships with explanatory variables. Usually, the spatial correlation of responses is modeled with a Gaussian random field based on a multivariate normal distribution. However, in practice, we encounter skewed responses, which are analyzed using skew-normal distributions. Closed skew-normal distribution is one of the extended families of skew-normal distributions, which has similar properties to normal distributions. This article presents a hierarchical Bayesian analysis based on a flexible subclass of closed skew-normal distributions. Given the time-consuming nature of Monte Carlo methods in hierarchical Bayes analysis, we have opted to use the variational Bayes approach to approximate the posterior distribution. This decision was made to expedite the analysis process without compromising the accuracy of our results. Then, the proposed model is implemented and analyzed based on the real earthquake data of Iran.

Stochastic Comparisons of Series and Parallel Systems Comprising Modified Additive Hazard Rate Model

Isaac Almasi

Adding parameters to a known distribution is a valuable way of constructing flexible families of distributions. In this paper, we introduce a new model, the modified additive hazard rate model, by replacing the additive hazard rate distribution in the general proportional add ratio model. Next, when two sets of random variables follow the modified additive hazard model, we establish stochastic comparisons between the series and parallel systems comprising these components.

Fuzzy Multiple Logistic Regression Under Fuzzy Random Errors

Hamid Reza Nili-Sani

In this paper, we consider the issue of data classification in which the response (dependent) variable is two (or multi) valued and the predictor (independent) variables are ordinary variables. The errors could be nonprecise and random. In this case, the response variable is also a fuzzy random variable. Based on this and logistic regression, we formulate a model and find the estimation of the coefficients using the least squares method. We will describe the results with an example of one independent random variable. Finally, we provide recurrence relations for the estimation of parameters. This relation can be used in machine learning and big data classification.

Survival Data Analysis Using Different Statistical Learning Methods

Kiomars Motarjem

Due to the volume and complexity of emerging data in survival analysis, it is necessary to use statistical learning methods in this field. These methods can estimate the probability of survival and the effect of various factors on the survival of patients. In this article, the performance of the Cox model as a common model in survival analysis is compared with compensation-based methods such as Cox Ridge and Cox Lasso, as well as statistical learning methods such as random survival forests and neural networks. The simulation results show that in linear conditions, the performance of the models mentioned above is similar to the Cox model. In non-linear conditions, methods such as Cox lasso, random survival forest, and neural networks perform better. Then, these models were evaluated in the analysis of the data of patients with atheromatous, and the results showed that when faced with data with a large number of explanatory variables, statistical learning approaches generally perform better than the classical survival model.

Dynamic Version of Weighted Cumulative Residual Extropy and Its Applications

Majid Hashempour

This paper introduces the dynamic weighted cumulative residual extropy criterion as a generalization of the weighted cumulative residual extropy criterion. The relationship of the proposed criterion with reliability criteria such as weighted mean residual lifetime, hazard rate function, and second-order conditional moment are studied. Also, characterization properties, upper and lower bounds, inequalities, and stochastic orders based on dynamic weighted cumulative residual extropy and the effect of linear transformation on it will be presented. Then, a non-parametric estimator based on the empirical method for the introduced criterion is given, and its asymptotic properties are studied. Finally, an application of the dynamic weighted cumulative residual extropy in selecting the appropriate data distribution on a real data set is discussed.

Piecewise Regression Model Based on Scale Mixture Normal Distribution

Farzane Hashemi

One of the most widely used statistical topics in research fields is regression problems. In these models, the basic assumption of model errors is their normality, which, in some cases, is different due to asymmetry features or break points in the data. Piecewise regression models have been widely used in various fields, and it is essential to detect the breakpoint. The break points in piecewise regression models are necessary to know when and how the pattern of the data structure changes. One of the major problems is that there is a heavy tail in these data, which has been solved by using some distributions that generalize the normal distribution. In this paper, the piecewise regression model will be investigated based on the scale mixture of the normal distribution. Also, this model will be compared with the standard piecewise regression model derived from normal errors.