Journal of Statistical Sciences

Record Linkage with Machine Learning Methods

Zahra Rezaei Ghahroodi

With the advent of big data in the last two decades, in order to exploit and use this type of data, the need to integrate databases for building a stronger evidence base for policy and service development is felt more than ever. Therefore, familiarity with the methodology of data linkage as one of the methods of data integration and the use of machine learning methods to facilitate the process of recording records is essential. In this paper, in addition to introducing the record linkage process and some related methods, machine learning algorithms are required to increase the speed of database integration, reduce costs and improve record linkage performance. In this paper, two databases of the Statistical Center of Iran and Social Security Organization are linked.

Stochastic Comparisons of Convolution of Independent Random Variables in the Scale Model

Ebrahim Nasiroleslami‎

This paper deals with some stochastic comparisons of convolution of random variables comprising scale variables. Sufficient conditions are established for these convolutions' likelihood ratio ordering and hazard rate order. The results established in this paper generalize some known results in the literature. Several examples are also presented for more illustrations.

Infinite Time Ruin Probability in the Risk Model of Excess Loss Reinsurance

Abouzar Bazyari

In this paper, the collective risk model of an insurance company with constant surplus initial and premium when the claims are distributed as Exponential distribution and process number of claims distributed as Poisson distribution is considered. It is supposed that the reinsurance is done based on excess loss, which in that insurance portfolio, the part of total premium is the share of the reinsurer. A general formula for computing the infinite time ruin probability in the excess loss reinsurance risk model is presented based on the classical ruin probability. The random variable of the total amount of reinsurer's insurer payment in the risk model of excess loss reinsurance is investigated and proposed explicit formulas for calculating the infinite time ruin probability in the risk model of excess loss reinsurance. Finally, the results are examined for Lindley and Exponential distributions with numerical data.

Clustering Based on Nonparanormal Graphical Mixture Models

Farzad Eskandari

Graphical mixture models provide a powerful tool to visually depict the conditional independence relationships between high-dimensional heterogeneous data. In the study of these models, the distribution of the mixture components is mostly considered multivariate normal with different covariance matrices. The resulting model is known as the Gaussian graphical mixture model. The nonparanormal graphical mixture model has been introduced by replacing the limiting normal assumption with a semiparametric Gaussian copula, which extends the nonparanormal graphical model and mixture models. This study proposes clustering based on the nonparanormal graphical mixture model with two forms of $ell_1$ penalty function (conventional and unconventional), and its performance is compared with the clustering method based on the Gaussian graphical mixture model. The results of the simulation study on normal and nonparanormal datasets in ideal and noisy settings, as well as the application to breast cancer data set, showed that the combination of the nonparanormal graphical mixture model and the penalty term depending on the mixing proportions, both in terms of cluster reconstruction and parameters estimation, is more accurate than the other model-based clustering methods.

Stochastic Comparison of (n-1)-out-of-n Systems from Multiple-Outlier Modified Proportional Hazard Rates Components in terms of Hazard Rate Order

Aliakbar hosseinzadeh

In this paper, we discuss the hazard rate order of (n-1)-out-of-n systems arising from two sets of independent multiple-outlier modified proportional hazard rates components. Under certain conditions on the parameters and the sub-majorization order between the sample size vectors, the hazard rate order between the (n-1)-out-of-n systems from multiple-outlier modified proportional hazard rates is established.

On Likelihood Ratio Ordering of Parallel Systems with Two Generalized Exponential Components

Mostafa Sattari

Consider two parallel systems with their component lifetimes following a generalized exponential distribution. In this paper, we introduce a region based on existing shape and scale parameters included in the distribution of one of the systems. If another parallel system's vector of scale parameters lies in that region, then the likelihood ratio ordering between the two systems holds. An extension of this result to the case when the lifetimes of components follow exponentiated Weibull distribution is also presented.

Modified Two-Stage Sampling Around the Mean of the First-Order Autoregressive Model

Eisa Mahmoudi

In this paper, a modified two-stage procedure in the Autoregressive model AR(1) is considered, which investigates the point and the interval estimation of the mean based on the least-squares estimator. The modified two-stage procedure is as effective as the best fixed-sample size procedure. In this regard, the significant properties of the procedure, including asymptotic risk efficiency, first-order efficiency, consistent, and asymptotic distribution of the mean, are established. Then, a Monte Carlo simulation study is deduced to investigate the modified two-stage procedure. The performance of estimators and confidence intervals are evaluated utilizing a simulation study. Finally, real-time series data is considered to illustrate the applicability of the modified two-stage procedure.

Comparison of Three Moment Methods of Parameter Estimation for FGM Copula in the Presence of Outlier

Hadi Jabbari

Paying attention to the copula function in order to model the structure of data dependence has become very common in recent decades. Three methods of estimation, moment method, mixture method, and copula moment, are considered to estimate the dependence parameter of copula function in the presence of outlier data. Although the moment method is an old method, sometimes this method leads to inaccurate estimation. Thus, two other moment-based methods are intended to improve that old method. The simulation study results showed that when we use copula moment and mixture moment for estimating the dependence parameter of copula function in the presence of outlier data, the obtained MSEs are smaller. Also, the copula moment method is the best estimate based on MSE. Finally, the obtained numerical results are used in a practical example.

Inverse Weibull-Poisson Distribution and Estimation of its Parameters in Type-II Censored Data

Parviz Nasiri

This paper presents the inverse Weibull-Poisson distribution to fit censored lifetime data. The parameters of scale, shape and failure rate are considered in terms of estimation and hypothesis testing, so the parameters are estimated under the type-II of censorship using the maximum likelihood and Bayesian methods. In Bayesian analysis, the parameters are estimated under different loss functions. The simulation section presents the symmetric confidence interval and HPD, and the estimators are compared using statistical criteria. Finally, the model's goodness of fit is evaluated using an actual data set.

Using Markov Latent Class Models in Estimating the Classification Error of Iranian Labor Flow Statistics

Lida Kalhori Nadrabadi

In countries where labor force surveys are based on rotation samples and partially standard sample units at different periods, the number of changing statuses can be estimated and presented as flow statistics. The response error is one of the essential non-sampling errors in labor force statistics. This error is doubled in flow statistics. Usually, the error of classifying flow statistics is estimated using the interview method, which is costly and complex. This paper presents the process of estimating flow statistics and appropriate models for calculating the classification error for it. Also, according to Iran's sample rotation pattern, each model's feasibility is examined. Finally, the Markov latent class model, assuming inequality of transition probabilities based on the rotation pattern of Iran for labor force samples, is introduced as a fit model for estimating classification error for flow statistics in Iran using the labour force survey data of 2019 and 2020.

Statistical Inference of Two-Parameter Weibull Distribution under Progressive Type-II Censoring with Random Removals

Maryam Sharafi

One of the most critical challenges in progressively Type-II censored data is determining the removal plan. It can be fixed or random so that is chosen according to a discrete probability distribution. Firstly, this paper introduces two discrete joint distributions for random removals, where the lifetimes follow the two-parameter Weibull distribution. The proposed scenarios are based on the normalized spacings of exponential progressively Type-II censored order statistics. The expected total test time has been obtained under the proposed approaches. The parameters estimation are derived using different estimation procedures as the maximum likelihood, maximum product spacing and least-squares methods. Next, the proposed random removal schemes are compared to the discrete uniform, the binomial, and fixed removal schemes via a Monte Carlo simulation study in terms of their biases; root means squared errors of estimators and their expected experiment times. The expected experiment time ratio is also discussed under progressive Type-II censoring to the complete sampling plan.

Comparison of Clustering High Dimensional Data by Random Projections Method and Some Common Methods of Dimensional Reduction

Mousa Golalizadeh

Nowadays, the observations in many scientific fields, including biological sciences, are often high dimensional, meaning the number of variables exceeds the number of samples. One of the problems in model-based clustering of these data types is the estimation of too many parameters. To overcome this problem, the dimension of data must be first reduced before clustering, which can be done through dimension reduction methods. In this context, a recent approach that is recently receiving more attention is the random Projections method. This method has been studied from theoretical and practical perspectives in this paper. Its superiority over some conventional approaches such as principal component analysis and variable selection method was shown in analyzing three real data sets.