Thailand Statistician

Performance of Phase II Control Chart for Location When One Parameter is Estimated in Terms of Run Length Distribution and Percentiles

2025-09-27T13:00:10+07:00

It is well known that the median run length measures chart performance better than the average run-length. Some authors have advocated more representative measures like the percentiles for assessing charts performance and called for an examination of the percentiles of the entire run-length distribution. An earlier work studied the percentiles of Shewhart chart when the process mean and variance are known (Case KK), later, another when the process mean and variance are unknown (Case UU). Here, we consider when only the process mean is unknown (Case UK) and when only the process variance is unknown (Case KU) by evaluating and plotting their exact run-length cumulative distribution functions for some reference samples (m) of size 5 at a given false alarm rate. We compare the results with those for Cases KK and UU. Unlike in Case UU, the cumulative distribution function curves for Case UK for small to moderate m are stochastically ordered relative to that of the geometric distribution and dominance is a function of δ, however, in line with Case UU, in Case KU, the curves cross that for the geometric distribution at some points and for at least m = 500 and
n = 5, the curves for Cases UK and KU converge with that for Cases KK and UU.

The Marshall-Olkin-Gompertz-Exponentiated Half Logistic-G Family of Distributions: Model, Properties and Applications

2025-09-27T13:55:13+07:00

In this work, a new family of distributions is introduced by combining Marshall-Olkin and Gompertz-exponentiated half logistic-G distributions. The new distribution is an infinite linear combination of the exponentiated-G family of distributions. Some of key properties including order statistics, ordinary moments, quantiles and moment generating function are derived. To estimate the
model parameters, the maximum likelihood method is also applied by means of Monte Carlo simulation study. Applications of the proposed family is demonstrated in many fields such as survival analysis and univariate data fitting. Empirical results show that the proposed models provide better fits than other well-known classes of distributions in different application fields.

A Non-Parametric Estimator of the Probability Weighted Moments for Large Datasets

2025-09-27T14:17:52+07:00

In this paper, we introduces a nonparametric median-of-means (MoM) estimator for Probability Weighted Moments (PWM) specifically designed for large datasets. Our approach draws inspiration from the data grouping method, a widely utilized technique in various domains including economics, hydrology, finance, and insurance. We establish the consistency and asymptotic normality of the
proposed estimator under reasonable assumptions regarding the increasing number of subgroups. Additionally, we present a novel approach for testing hypotheses related to Probability Weighted Moments (PWM) using the Empirical Likelihood method (EL) specifically tailored for the median. Notably, our method circumvents the need for prior estimation of the variance structure associated
with the estimator, a task that can be challenging and prone to inaccuracies. We conducted numerical simulations to assess the performance of our proposed estimator. The results clearly indicate that our estimator showcases remarkable robustness, particularly when confronted with outliers.

Acceptance Sampling Plans Based on Truncated Life Tests for the Nadarajah-Haghighi Distribution

2025-09-27T14:48:06+07:00

In this paper, we develop a time truncated single acceptance sampling plan when the lifetime follows the Nadarajah-Haghighi distribution. We obtain the minimum sample size necessary to ensure a certain mean lifetime for various selected acceptance numbers, consumer’s confidence levels and values of the ratio of the experimental time to the specified mean lifetime. The operating characteristic function and the associated producers risks are also studied. A numerical example is provided to illustrate some results of the paper.

An Extended Approach to Test of Independence between Error and Covariates under Nonparametric Regression Model

2025-09-27T15:05:42+07:00

In 2014, Bergsma et al. (2014) proposed a generalized measure of association τ∗ as an extension of widely used Kendall’s τ. Later, in testing of independence between error and covariate, under nonparametric regression model Y = m(X) + ϵ, with unknown regression function m and observation error ϵ, test statistic tailored on τ ∗ was suggested by Dhar et al. (2018). In this article, we develop a test, constructed on further extension of τ ∗, considering the ordered X and the third order difference of Y with an motive to address the same issue of independence. We deduce the asymptotic distributions of test statistics using the theory of degenerate U-statistics. Moreover, we unravel the power of the proposed tests using Le Cam’s concept of contiguous alternatives. A couple of simulated examples on normal and non normal distribution are furnished. Also, the performance of the test statistics is honed through a real data analysis.

A Novel Distribution with Upside-down Bathtub Shape Hazard Rate: Properties, Estimation and Applications

2025-09-27T15:18:41+07:00

In this communication, we introduce a new statistical model and study its various mathematical properties. The expressions for hazard rate, reversed hazard rate, and odd functions are provided. We explore the asymptotic behaviours of the density and hazard functions of the newly proposed model. Further, moments, median, quantile, and mode are obtained. The cumulative distribution and density functions of the general kth order statistic are provided. Sufficient conditions, under which the likelihood ratio order between two different inverse generalized linear failure rate (IGLFR) distributed random variables holds, are derived. In addition to these results, we introduce several estimates for the parameters of IGLFR distribution. The maximum likelihood and maximum product spacing estimates are proposed. Bayes estimates are calculated with respect to the squared error loss function. Further, asymptotic confidence and Bayesian credible intervals are obtained. To observe performance of the proposed estimates, we carry out a Monte Carlo simulation using R software. Finally, two real life data sets are considered for the purpose of illustration.

Parameter Estimation for Generalized Random Coefficient in the Linear Mixed Models

2025-09-27T15:40:09+07:00

The analysis of longitudinal data, comprising repeated measurements of the same individuals over time, requires models with a random effects because traditional linear regression is not suitable and makes the strong assumption that the measurements are independent, which is often unrealistic for longitudinal data. However, values repeatedly measured in the same individual are usually correlated, and ignoring the correlation between repeated measurements may lead to biased estimates as well as invalid P-values and confidence intervals. Therefore, careful consideration is needed to enable valid inference of covariate effects on longitudinal responses. In this regard, Residual Maximum Likelihood (REML) analysis is the most widely used method to estimate parameters. This method is based on the assumption that there is no correlation between the random effects and the error term (or residual effects). Hence, it is unclear if the failure to meet this assumption will affect the conclusions when conducting a real datasets study. In the present article, we propose Conditional Least Squares (CLS) and Weighted Conditional Least-Squares (WCLS) methods for estimating the model parameters, considering the presence of the correlation between the random effects and the error term. The choice of these methods is motivated by the fact that it is not necessary to specify the distributions of random effects and/or errors. These methods are illustrated via simulation studies that were performed with different sample sizes and different parameter values. In addition, we compared these methods with traditional estimators. This comparison is made by utilizing the ratio of their mean square error. Our results highlight that the weighted conditional least-squares estimator is efficient and attractive compared to the restricted maximum likelihood when random effects are permitted to be correlated with the error term. Also, real data analysis is conducted to confirm the advantages of the improved method.

Programs for Performing the A Priori Procedure in Skew Normal Contexts

2025-09-27T15:57:26+07:00

Despite being only a few years old, there is already a reasonably large literature on the a priori procedure, designed to aid researchers in determining sample sizes needed to ensure that sample statistics to be obtained provide good estimates of corresponding population parameters. One of the directions this literature has taken has been to expand the typical assumption of normally distributed data to include skew normally distributed data. In turn, the expansion allows the use of locations, or differences in locations; as opposed merely to means, or differences in means. Another direction the expansion has taken is to consider correlation coefficients. However, there is considerable mathematics associated with both expansions that might not be accessible to most substantive researchers. Therefore, the present goal is to provide links to free and user-friendly programs so that even mathematically unsophisticated substantive researchers can perform the calculations.

Unpacking the Opinion of Taiwanese on COVID-19 Policies: Going Beyond Simple Aggregate Satisfaction

2025-09-27T16:16:08+07:00

This research examines Taiwans public health behaviors, attitudes, and policy opinions regarding COVID-19 policies during the pandemic, taking South Korea as a comparison. We assess how public approval of COVID-19 policies correlates with partisanship, voting behavior, and confidence in government and healthcare during the pandemic. We focus on three major issues: 1) Identifying
key factors associated with confidence in the governments pandemic response. 2) Investigating the connection between presidential voting behavior and pandemic policies. 3) Evaluating respondents confidence in specific policies and the healthcare system during the pandemic. We collect samples from the Taiwan Social Change Survey (TSCS) conducted between September and December 2021,
obtaining 1,260 complete face-to-face interview responses. We use Goodman and Kruskals lambda to measure the correlation between each policy and the respondents responses. Using decision tree learning and considering the region as a random effect in the mixed-effects logistic regression, we examine the relationships between political outcome variables and their explanatory counterparts. Our study reveals that political factors and demographic characteristics significantly shape public confidence in Taiwans governmental and healthcare responses to COVID-19. The majority of Taiwanese people, irrespective of voting preferences, generally support measures such as mobile phone surveillance to track infected individuals, border closures, mandatory face masks, and isolation of those with the disease to curb COVID-19 transmission. Additionally, the closure of educational institutions as a public health measure enhances public confidence in the government.

Unleashing the Potential: Exploring Cashless Investment Strategies through Short-Selling Techniques and Modern Portfolio Theory

2025-09-27T16:41:21+07:00

This research paper investigates cashless investment strategies using the short-selling technique and the principles of Markowitz’s Portfolio Theory and Modern Portfolio Theory (MPT). The study examines the linear relationship between securities in the SET50 index group using correlation coefficient and beta coefficient values and creates a portfolio of two stocks with a negative correlation.
Using short-selling techniques, we develop a model to identify suitable investment portfolios for cashless investment strategies and analyze five years of daily closing prices of 41 securities in the SET50 index group. We use the cross-exponential moving average 2 lines (Cross EMA) to set buysell signals and evaluate performance using Shape Ratio, Success Rate, and Net Margin Profit. The
results show 72 suitable investment portfolios for cashless investment strategies using short-selling techniques, ranked by their net margin profit, providing investors with valuable insights and recommendations. We discuss the implications of our findings and outline areas for future research while acknowledging the study’s limitations. This study demonstrates the potential of cashless investment
strategies using the short-selling technique and the value of applying Markowitz’s Portfolio Theory and MPT to portfolio construction.

Inventory Model for Non-Instantaneous Deteriorating Items with Stock-Dependent Demand and Preservation Technology Investment

2025-09-27T17:10:04+07:00

The paper considers an inventory system for non-instantaneous deteriorating items when demand is dependent on the stock level. Time dependent deterioration and partial waiting time dependent backlogging rates are assumed to model the inventory system. Deterioration, although non-instantaneous, is not desirable. Hence some technique that can reduce deterioration is very necessary. However, that will involve a certain amount of cost. This preservation technology investment and its effect on the inventory decision are considered in this paper. The optimum cost function is formulated, and its convexity is proven. The study is supported by hypothetical illustrations that suggest that the minimum cost decreases owing to the preservation technology adopted. Managerial aspects, future scope and the limitations of the study are also discussed alongside the findings and conclusions. It is concluded that although preservation technology involves certain amount of investment, the total cost is reduced by its application.

Stratified Folded Ranked Set Sampling with Perfect Ranking

2025-09-27T18:55:38+07:00

This study introduces the Stratified Folded Ranked Set Sampling with Perfect Ranking (SFRSS) method, a novel approach to enhance population mean estimation. SFRSS integrates stratification and folding techniques within the framework of Ranked Set Sampling (RSS), addressing inefficiencies in conventional methods, particularly under symmetric distribution assumptions. The unbiasedness of the SFRSS estimator is established, and its variance is shown to be lower compared to Simple Random Sampling (SRS), Stratified Simple Random Sampling (SSRS), and Stratified Ranked Set Sampling (SRSS). Simulation studies conducted across Uniform, Normal, and Student-t distributions demonstrate the superior efficiency of SFRSS, particularly for heavy-tailed distributions, where ranking and folding significantly reduce variance. The findings highlight SFRSS as a robust alternative for stratified sampling, providing practical benefits in scenarios where population symmetry and stratification play a key role.

Bayesian Estimation and Prediction for Zero-Inflated Discrete Weibull Distribution

2025-09-27T19:04:33+07:00

This paper proposes the Bayesian estimation of the zero-inflated discrete Weibull distribution assuming three prior distributions, namely Beta-Uniform-Uniform prior, Beta-Jeffreys’ rule prior, and Beta-Beta-Gamma prior. It is commonly known that there is no compact form for the Bayes estimators. The Bayesian estimate of the model parameters has been performed through the random walk Metropolis algorithm. The Bayes estimates of the unknown parameters and the credible interval construction are established based on the generated samples. Moreover, the maximum likelihood estimation is considered, as well as the confidence interval estimation for the model parameters has been performed through normal approximation. The performance of the Bayes estimators has also been compared with the classical estimators through the Monte Carlo simulation study. Further, the Bayesian prediction of a future observation is proposed under the three prior distributions. The posterior predictive distribution of a future observation cannot be evaluated analytically. To obtain the estimate of future sample, the Metropolis-Hastings algorithm is used. Two datasets have been analyzed to show how the proposed model and the method work in practice.

One-Period Coupon Bond Valuation Using the Variance Gamma Model

2025-09-27T19:36:13+07:00

Bonds issued by companies all have coupons. One-period coupons are issued once and are paid together with the principal amount of the bonds at maturity. Many studies on coupon bond valuations have been carried out using a structural model approach. The model involves the assets of the company. The asset model that is often used is the Black-Scholes-Merton model, which assumes in asset returns with a normal distribution. But in reality, the financial data is not normally distributed and indicates the presence of heavy tails and excess kurtosis. In this study, bond valuation was carried out using the variance gamma (VG) model approach. An empirical study was conducted on bond data from one of the banking companies in Indonesia, namely the Continuous Bond III of Bank CIMB Niaga Phase I 2019 Series B. The bonds were issued on December 19, 2019, with a face value of IDR 1,066,000,000,000 and a maturity date of December 19, 2022. The VG asset model in this study has a mean absolute percentage error (MAPE) value of 1.59 % which gives the conclusion that the asset model is very accurate. The bond valuation shows an expected equity value of IDR 271,071,400,000,000, an expected liability of IDR 31,359,700,000,000, and a default probability of 0. This result indicates that PT Bank CIMB Niaga Tbk is able to fulfill its obligations when the bonds mature.

Canonical Correlation Analysis on Physicochemical Data and Proximate Data in the Case of Goat Milk Yoghurt Mixed with Basella spp. Fruit Powder

2025-09-27T20:14:20+07:00

This study aims to examine the relationships between two datasets of an experiment on goat milk yoghurt mixed with Basella spp. fruit powder. We use canonical correlation in investigating the relationships between physicochemical variables: pH, lactic acid, viscosity, L* color, a* color, b* color and proximate variables: protein, fat, ash, fibre, humidity, carbohydrate. Our results are obvious that some variables of two data have high canonical correlations, that is, {viscosity, L* color, a* color, b* color} and {fibre, humidity, carbohydrate}, which are used for constructing canonical variates. We obtain that the first canonical variates explain the proportion of variability of about 67.43% with high canonical correlation 0.902, whereas the second and third canonical variates explain the proportion of variability of about 21.52% and 11.05%, respectively. Consequently, a squared canonical correlation of the first canonical variates is high (0.8136), that is, 81.36% of the variation in the first physicochemical canonical variable is explained by the variation in the first proximate canonical variable. These results are very useful for designing our next experiment in reducing the complexity of data collection and extravagant expense. We can also develop nutrition of our yoghurt in the future.