نشریه اندیشه آماری
سال بیست و سوم شماره 1 (پیاپی 45، بهار و تابستان 1397)

Due to the economical restrictions, blood centers need to detect approaches to improve their efficiency. Improving efficiency in the collection and processing of blood products at blood centers can decrease costs while assuring that demand for blood products is happened. One of these approaches can be allocating resources more effectively Study designs and This study uses data envelopment analysis (DEA) to evaluate the efficiency of 31 Provincial units of Iranian Blood Transfusion Organizations (IBTO) to determine to what extent efficiency can be improved. Input variables included Number of blood collection and processing centers (BCPCs), Number of blood collection centers (BCCs), Number of mobile teams(MTs), Population density of province (population per square mile), Number of donor beds and Area of blood transfusion center (BTC). And the output variables were produced Units of Packed cells (PCs), FreshFrozen Plasma (FFP) and platelets (PLTs) for use inside province. Efficiency grades were computed
with DEA linear programming techniques and Provincial units of IBTO characteristics that important affect efficiency determined by principle components analysis (PCA). For two consecutive years, with Number of BCPCs, BCCs, MTs, Population density of province, Number of donor beds and area of BTC as inputs, technical efficiency was (range, 0.39- 1). for two consecutive years, 22 Provincial units of IBTO were efficient (efficiency, = 100%), and 9 were inefficient. Otherwise based on PCA results, Efficiency was mainly affected by Number of BCPCs, BCCs, MTs, Efficiency did not directly relate to Population density of province, Number of donor beds and area of BTC. Conclusions: The major reason of inefficiency was excess allocating resulting from a suboptimal combination of Number of BCPCs, BCCs, and MTs. Further research is required to manage factors affecting efficiency, such as the variation of demand inside and outside the provinces.

In this paper we explain a necessary and sufficent condition for independence between any arbitrary statistics with sufficient statistics which is also maximum likelihood estimator in a general exponential family with location and scale parameter namely generalized normal distribution. At the end, it is shown that the converse is true except in the asymptotic cases.

In this paper some properties of Beta - x family are discussed and a member of the family,the beta– normal distribution, is studied in detail .One real data set are used to illustrate the applications of the beta-normal distribution and compare that to gamma - normal and Birnbaum-Saunders distriboutions .

Often in practice the data on the mortality of a living unit with a correlation of location of the observations in space are studied. One of the most important issues in this type of analysis of survival data with spatial dependence is estimation of the parameters and prediction of the unknown values in known sites based on observations vector. In this paper to analyze this type of survival data, Cox regression model with piecewise exponential function is used as a hazard and spatial dependence as a Gaussian random eld and as a latent variable is added to the model. Because there is no closed form for the posterior distribution and the full conditional distributions, also long computing for Markov Chain Monte Carlo of algorithms, to analyze the model is used the approximate Bayesian approach. In a practical example is presented how to implement an approximate Bayesian approach.

Many studies have been done on the relationship between self-regulated learning and academic performance of students whose results have also conflicting. This study aims at a methods of systematic review, called meta-analysis, to analysis the results of preceding studies carried out in Iran about  role of Self-regulated learning on learners' academic performance in the past decade. Studies investigating the relationship between self-regulated learning and academic achievement have been conducted mainly in the form of statistical classical models, while the nature of these variables and the relationship between them are fuzzy too. It is suitable, therefore to imploy a fuzzy method to analysis such data. fifty accomplished researches about role of Self-regulated learning on learners' academic performance, 31 researches were chosen for fuzzy meta-analysis. Findings shown that there is a meaningful relationship between  Self-regulated learning and Learners' academic performance and  Self-regulated learning could explain 4-17 percent  of  variance in academic performance. The results can be inclusive education program planning and effective learning them.

The normal distribution plays an important role in statistical analysis. However, a researcher may also wish to construct another symmetric distributions which fit the data better than the normal distribution. For this purpose, more flexible distributions have been introduced. In this thesis, we introduce some of such distributions. We first introduce the slash distribution as a family of mixed-normal scale distributions. The slash distribution can be used instead of normal distribution, in many situations. We also introduce Skew-slash. Then, a new modified-slash distribution is discussed. We also study a modified Skew-slash distribution. Some properties of the distributions discussed in this thesis are given. In particular, we present the stochastic representations, density functions and moments of the distributions.

The Borel-Cantelli Lemma is very important in the probability theory‎. ‎In this paper‎, ‎we first describe the general case of the Borel-Cantelli Lemma‎. ‎The first part of this lemma‎, ‎assuming convergence and the second part includes divergence and independence assumptions‎. ‎In the following‎, ‎we have brought generalizations of the first and second part of this lemma‎. ‎In most generalizations of part II‎, ‎the condition of independence‎, ‎pairwise independence‎, ‎weakening and elimination of the condition of independence have been investigated‎.

The goal of this research is finding the best method of equating data in item response theory. For that purpose, we are using the data of Ph.D test in  Statistic major for two consecutive years 92 and 93. For analyzing, we are specificly using the tests of Statistics major which includes 45 questions. Parameters of test and ability of individuals are considered according to the three parametrs model and by using the MULTILOG software. In this study, we are using the Mean-Mean method, Mean-Sigma, Haebara, and Stocking-Lord methods by considering the unequal groups with Anchor-test design, and we are using the root mean square error for choosing the optimal solution. The results of this study show that the methods under characteristic curve are more accurate.

In this paper, we  Hurdle  regression  models  for count responses. A full likelihood-based approach is used to yield maximum likelihood estimates of the model parameters.  Our model is applied to a insurance data set. In this data set, the  number of  claim  count is considered as count response.  Different count regression models  are used to model these  power series responses. Some of  these count  regression models are Hurdle Poisson and Hurdle Negative Binomial regression model.

When the observations reflect a multimodal‎, ‎asymmetric or truncated construction or a combination of them‎, ‎using usual unimodal and symmetric distributions leads to misleading results‎. ‎Therefore‎, ‎distributions with ability of modeling skewness‎, ‎multimodality and truncation have been in the core of interest in statistical literature‎, ‎always‎. ‎There are different methods to contract a distribution with these abilities‎, ‎which using the weighted distribution is one of these methods‎. ‎In this paper‎, ‎it is shown that by using a weight function one can create such desired abilities in the corresponding weighted distribution.