Thailand Statistician

Exponential Estimator in the Absence and Presence of Non-Response using Double Sampling

Sunil Kumar — 2025-03-29

This paper suggests modified exponential estimators for estimating the population mean of the study variable by utilizing the information on the auxiliary variable in double sampling scheme under two situations: i) when there is full information available on the study as well as auxiliary variable, ii) when there is a non-response on the study as well as auxiliary variable. A theoretical study is
carried out when the population mean of the auxiliary variable is not known. The expressions of the biases and mean squared errors (MSE’s) have been obtained under large sample approximation under double sampling. Theoretical comparison of proposed estimator with already existing estimators has been studied and the favourable conditions have been obtained for which proposed estimator is better than existing estimators. An empirical study has also been done in the support of theoretical results.
Moreover, these theoretical findings are supported by simulation study.

Mean Estimation in Presence of Measurement Errors Using Log Type Estimators

Shashi Bhushan — 2025-03-29

This article introduces some log type class of mean estimators in the case of measurement errors (ME) using simple random sampling (SRS). The mean square error of the proposed estimators is obtained when data on both the study and auxiliary variables are commingled with ME. The performance of the proposed estimators is compared with the existing estimators and the efficiency conditions are derived. Further, the performance of the proposed estimators is illustrated through numerical and simulation studies using some real and artificially generated populations. The results of numerical and simulation studies show that the proposed estimators dominate the usual mean estimator, classical ratio and product estimators.

Log Type Estimators Using Multi-Auxiliary Information Under Ranked Set Sampling

Shashi Bhushan — 2025-03-29

It is well known that the relevant utilization of auxiliary information associated with the auxiliary variable helps to enhance the efficiency of the estimates. Therefore, we introduce some log
type estimators based on multi-auxiliary information under ranked set sampling. The mean square error (MSE) of the suggested estimators is derived to the first order approximation. The efficiency conditions are obtained by comparing the MSE of the suggested estimators with the MSE of the contemporary estimators. Further, numerical and simulation studies are conducted over real and artificially generated populations to support the theoretical results. The empirical results show that the suggested estimators perform better than the usual mean estimator, classical ratio estimator, AbuDayyeh et al. (2009) estimator, Mehta and Mandowara (2014) estimator, Khan and Shabbir (2016) estimator and Khan et al. (2019) estimator

Bayesian Inference for Nadarajah-Haghighi Distribution Under Progressively Type-II Censored Data

Reza Azimi — 2025-03-29

In this article, the problem of estimating parameters, reliability function, and hazard function
of the Nadarajah-Haghighi distribution under progressively type-II censored samples are studied.
The maximum likelihood and Bayesian estimators under squared error, LINEX, and general entropy loss functions are derived for parameters and some survival time parameters namely reliability and hazard functions. We used Lindley’s approximation to obtain the Bayesian estimators. Asymptotic confidence intervals, for unknown parameters, are constructed using the observed Fisher information
matrix. A numerical example using the real data set is presented to illustrate the proposed methods. Monte Carlo simulation study is conducted to compare the performance of the estimators in terms of their mean square error. The Monte Carlo simulation analysis shows that in most cases, the Bayesian method have a better performance than the standard maximum likelihood method.

Unleashing Competitive Potential: An Intuitive Fuzzy Logic Approach for Assessing Securities Firm Performance

Wichai Witayakiattilerd — 2025-03-29

This study presents a mathematical model designed to assess the competitive advantage of securities firms by analyzing eight key factors: service quality, product introduction, product design tailored to user needs, commission fees with and without investment advisors, organizational structure, return on assets (ROA), and return on equity (ROE). Utilizing fuzzy logic and a weighted average approach, the model calculates an overall competitive advantage score for each firm, facilitating a ranking system that highlights the firms with the highest competitive potential. This ranking system not only reflects a firm’s strategic and operational strengths but also aligns with external benchmarks from prominent financial websites, offering a comprehensive view of market positioning. The study’s
conclusions emphasize the model’s utility in guiding firms to enhance their strategic planning and improve service offerings, while also serving as a valuable tool for investors and policymakers to understand the competitive landscape within Thailand’s securities sector. By providing insights into areas of potential growth and market opportunities, this research aids firms in refining their competitive strategies and bolstering their market presence.

The Unit Weibull Regression Model with Variable Shape Parameter

Lucas David Ribeiro-Reis — 2025-03-29

In this paper a generalization of the unit Weibull regression model is introduced. Here, both the median and the shape parameter are modelled through covariates. The parameters are estimated by maximum likelihood. Analytical expressions for the score vector and the Fisher’s observed information matrix are demonstrated. A simulation study is performed to show the consistency of the maximum likelihood estimators. Finally two applications to real data from Brazil are considered. These applications show the usefulness of the proposed model.

On Some Aspects of Exponential Intervened Geometric Distribution

Jayakumar Kuttan Pillai — 2025-03-29

In this paper, we study different aspects of Exponential Intervened Geometric (EIG) distribution. EIG distribution arises as the distribution of random minimum and is a generalization of extended exponential distribution. The shape properties of the probability density function and hazard rate function of EIG are studied, along with structural properties such as moments, moment generating
function, skewness and kurtosis, mean deviation about mean and median. Expression for various reliability measures corresponding to EIG distribution are derived along with stochastic ordering property. Expression for quantiles are obtained and random number generation is discussed. The distributions of order statistics are derived and limit distributions of sample extrema are obtained. Four
characterizations of EIG distribution are proved. The parameters of EIG are estimated through the method of maximum likelihood (ML) and a simulation study is conducted to show the performance of ML estimates. The existence and uniqueness of ML estimates are proved. The EIG model is fitted to a real data set and is showed that the model performs better as compared to ten competitive models. Also, the adequacy of the model for the data set is established using parametric bootstrap approach.

A Pseudo Estimation of Variance using Prior Information with Unknown Shape Parameter: A Study in Normal Case

Shreya Bhunia — 2025-03-29

In this article, a variant of Bayes estimate for variance parameter from normal distribution is
investigated under weighted squared error loss function. For this purpose, inverse gamma conjugate
prior distribution with an unknown hyper-parameter is considered. Further, empirical Bayes approach
is used to estimate that unknown hyper-parameter from the marginal likelihood equation. We consider some numerical iterative procedure to approximate the hyper-parameter and in this context, the Bayes estimator may be called a pseudo empirical Bayes estimator. To study the performance of the estimator, the integrated risk and bias performance are carried out through an extensive simulation. It is seen that the estimator performs well in accordance with the integrated risk values. To reduce the bias of the estimator, jackknife resampling technique is used further. Finally, the efficiency of the proposed estimator is studied using two real-life datasets and it is found to be satisfactory as they produce lower posterior risk values.

Akaike Information Criterion for a Simultaneous Equations Model using Unbiased Estimator

Warangkhana Riansut — 2025-03-29

This article introduces a novel approach in the fields of econometrics and statistics. A new criterion for selection in a simultaneous equation model (SEM) based on the Akaike information criterion using an unbiased estimator (UE) called SAIC(UE). This innovative method replaces the traditional maximum likelihood estimator (MLE) with UE, offering a fresh perspective on model selection. A comparison of the effectiveness of SAIC(UE) with SAIC (Keerativibool et al. 2011) by the extensive simulation study and the observed L₂ performance found that SAIC(UE) is the best criterion because it had the maximum percentage of correct model selections, the maximum average of the observed L₂ performance, and the minimum standard deviation of the observed L₂ performance in all situations of the simulation study. SAIC has a more negative bias than SAIC(UE) because the contemporaneous covariance matrix of error terms obtained from MLE is smaller than that obtained from UE or underestimated. This reason causes the correct order of the model from SAIC to be less than SAIC(UE).

Optimal Fenton Process Using the Modified Taguchi Approach

Rajyalakshmi Kottapalli — 2025-03-29

Unavoidable scatter in repeated test results can be due to influence of unknown process variables (if any) and measurement errors to a certain extent. Modified Taguchi approach recommends few tests as per the orthogonal array and provides the range of estimates for combinations among the levels of process variables. It identifies optimal process variables and demands additional experimentation for confirmation (if necessary). One of the widely applied advanced oxidation processes (namely, Fenton oxidation) utilizes ferrous iron and hydrogen peroxide under acidic conditions to produce hydroxyl radical. This article presents optimal process variables for COD and decolorization efficiency of Fenton oxidation adopting multi-objective optimization. Empirical relations are presented for COD and decolorization efficiency in terms of Fenton oxidation process variables. Comparative study indicates reasonably good agreement between empirical relations and test results.

Comparison of Methods Used for Estimation of Number of Factors for High Dimensional Factor-Augmented Functional-Coefficient Forecasting Models

Jiraroj Tosasukul — 2025-03-29

Estimating the number of common factors is challenging particularly in nonparametric high dimensional time series model where the dimensions are bigger than the sample size being considered. In this article, we study the different techniques for determining the number of common factors which are considered in the context of reducing dimensions in the factor-augmented functional-coefficient model with high dimensions. The concept behind all the estimate techniques that are used in this research is based on the evaluation of eigenvalues derived from the correlation and covariance matrix. The performance of the estimators is compared using percentages and the average of the estimated factor numbers. The results of the simulation experiments showed that the growth rates of residual variances technique exhibits outstanding performance compared to other methods in situations when the dimension is greater than or equal to the large sample size. Nevertheless, both modified eigenvalues thresholding and the eigenvalue difference criterion techniques provide better results in comparison to other methods in cases when the dimension is smaller than the small sample size. This study employed an empirical approach, using an actual dataset of Australia's quarterly consumer price index (CPI) to provide evidence for the estimating techniques used for determining the factor number.

Construction of Quality Regions Based on Consumer’s and Producer’s Risks for Two Sided Group Chain Sampling Plan with Binomial Distribution

Waqar Hafeez — 2025-03-29

Acceptance sampling is a technique that is used to make decision about the lot under inspection. Based on representative sample the decision is made about the whole lot that is under inspection. Mostly existing plans consider consumer’s risk and they do not care about producer loss. This study will consider consumer’s risk as well as producer’s risk and provide a criterion to satisfy both parties at the same time. A two sided group chain sampling plan (TSGChSP) is used in this paper. Based on both risks probability of lot acceptance, L(p) is determined. Four different quality regions are estimated for specified values of producer’s and consumer’s risks. By satisfying the specified design parameters, it is assessed that as the value of design parameters increases, the proportion of defectives decreases such as g, r, i, j, and . In comparison TSGChSP is compared with existing Bayesian two sided group chain sampling plan (BTSGChSP). If both plans are applied under similar environments, then the results explain that the TSGChSP produces a lower proportion of defectives than the BTSGChSP. Hence, we suggest that TSGChSP is better substitute for lot inspection in the manufacturing industry, particularly for those working with destructive testing of high-quality products.

Enhancement of Systematic Sampling with Incorporation from Consecutive Sampling; Systematic Sampling with Consecutive Approach

Mohamad Adam Bujang — 2025-03-29

This study develops a newly modified systematic sampling that allows of recruiting more than one-unit subject in each interval selection. The newly modified systematic sampling is an incorporation from both consecutive sampling and systematic sampling. Firstly, the terms and selection of the newly modified systematic sampling is identified when the sample selection derived from the new sampling technique is able to generate an unbiased estimator for the population mean. Then, the evaluation from simulation works and a real-life dataset was applied to support the evidence of sampling efficiency and consistency. This newly modified systematic sampling is named as “Systematic Sampling with Consecutive Approach” or in short SSC is indeed an innovation in sampling by the virtue of combining two sampling techniques, consecutive sampling and systematic sampling.

Estimation of Population Mean in the Presence of Non-Response for Time-Based Surveys

Sanjay Kumar — 2025-03-29

An estimate of the location of a distribution is the most fundamental type of inference about a population. Consequently, obtaining more accurate estimators of the population mean of interest is essential in every statistical estimating technique. Survey statisticians most often use priori information at an estimation stage to form an estimator for estimating parameters. In this study, we suggested an estimator for the population mean in the presence of non-response utilizing information from the past surveys along with information available from the current surveys in the form of a hybrid exponentially weighted moving average. We obtained the expressions of the suggested estimator's mean and variance and established the mathematical conditions to demonstrate the efficiency of the suggested estimator. We supported the theoretical outcomes with the help of a simulation study and a real-life example. The results show that the utilization of information from past surveys along with the current surveys improves the efficiency of the suggested estimator. For example: in the simulation study, for a sample size n=(50, 100) at the non-response rate (=0.20, 0.15) and weights (0.10) and (0.15) to the current and past observations, variances of the suggested estimator were 0.000683, 0.000559; 0.000635, 0.000309), which were less than (0.020944, 0.009756; 0.020024, 0.008920) of the existing Hansen and Hurwitz (1946) estimator. Similarly, in the empirical study of the real-life dataset, for the sample number (=6) having size n (=6) at (=0.30, 0.25), (=0.10) and (=0.15), variances of the suggested estimator were (0.74, 0.71), which were significantly less than (95.12, 90.65) of the existing Hansen and Hurwitz (1946) estimator. The suggested work is limited to the homogeneous population only.

Bayesian Hybrid Tripartite Randomized Response Technique

Adetola A. Adediran — 2025-03-29

The response bias is a fundamental problem with wide range of tendencies for participants to respond inaccurately or falsely to questions. These biases are prevalent in research involving sensitive questions which have a large impact on the validity of surveys, In the quest to reduce bias in surveys, Warner (1965) proposed the Randomized Response Technique (RRT) which has undergone so many improvements/modifications. To further reduce these biases, this paper proposed a Bayesian estimation of Hybrid Tripartite Randomized Response Technique (BHTRRT) by using prior information to improve existing works on Hybrid Tripartite Randomized Response Technique (HTRRT). The Bayesian approach has emerged as a strong competitor to the traditional classical approach to randomized techniques; hence, its approach to HTRRT. The Bayesian approach accommodates other intrinsic parameter constraints in the posterior to improve statistical precision. The effectiveness of the proposed method was examined by conducting simulation study and compares the performance of the proposed estimator with conventional ones. Result showed that the proposed technique is more efficient and reduce response bias in survey better than the conventional ones.