Thailand Statistician

Frailty-Based Competing Risks Model for the Analysis of Events in Transition to Adulthood

Jayanta Deb Deb — 2024-06-29

The analysis of clustered time-to-event data is carried out using random effects models, popularly known as frailty models in the literature of event history analysis. The present work demonstrates an application of competing risks frailty model for analysing adulthood transitions that are clustered into geographical regions. Observations from the same cluster are assumed to be correlated because these usually share certain unobserved characteristics. Ignoring such correlations may lead to incorrect standard errors of the estimates of parameters of interest. Besides making the comparisons between usual competing risks model and competing risks model with frailty for analysing geographically clustered time-to-event data, important demographic and socio-economic factors that may affect the duration of transition to adulthood events namely: transition from leaving study to work and/or marriage of Indian youths are also reported in this paper. The data from the study ”The Youth in India: Situation and Needs 2006-2007” which was implemented by the International Institute for Population Sciences, Mumbai and the Population Council, New Delhi (IIPS and PC 2010), is used. The results of the analysis highlight the significant transition differentials among Indian youths by gender, place of
residence, religion, caste, wealth quintile, among others. We found that after leaving study men join the work much earlier than women, and prefer to postpone their marriage. But women have higher likelihood of entering into marriage early compared to men. Rural residents have significantly higher likelihood of joining work and lower likelihood of entering into marriage compared to their urban counterparts at their early age. Wealth quintile has been observed to have a mild or no significant effect on the hazards of adulthood transitions.

Comparing Some Iterative Methods of Parameter Estimation for Progressively Censored Lomax Data

Amal Helu — 2024-06-29

Based on Progressively Type-II censored samples, the maximum likelihood estimator, the uniformly minimum variance unbiased estimator (UMV U), and the Bayes estimators for the shape parameter and the hazard function of the Lomax model are derived. The Bayesian estimators are obtained based on symmetric (squared error, absolute difference, and logarithmic loss functions) and
asymmetric (LINEX, General Entropy, and Logarithmic) loss functions. A real data example consists of data from Iowa 65+ Rural Health Study (RHS) is used to illustrate the proposed methods

Robust Inference for the Skew Normal Regression Model Under Type II Censoring

Iklim Gedik Balay — 2024-06-29

In this paper, we concentrate on statistical inference for the regression model with skew normal (SN) distributed error terms under type-II censoring. Iteratively reweighting algorithm (IRA) is used for computing maximum likelihood (ML) estimates of the model parameters, see Arslan (2009). We also use the non-iterative modified maximum likelihood (MML) methodology to obtain the explicit
estimators of the model parameters, see Tiku (1967). Additionally, confidence intervals for the model parameters are constructed based on the proposed estimators. Monte Carlo simulation study is used to compare the efficiencies of the ML and MML estimators, and also the performances of the corresponding confidence intervals.

On the New Mixture Sushila and Rayleigh Distributions: Mathematical Properties and Application

Ibrahim Abdullahi — 2024-06-29

In this study, a newly proposed mixture of Sushila and Rayleigh distributions is presented, introducing a right-skewed pattern in the distribution of lifetime data. The fundamental properties of the distribution have been derived. Parameter estimation for the new mixture of Rayleigh and Sushila distributions is conducted using the maximum likelihood method. The application of this new mixture in modeling remission times for bladder cancer is explored. Results indicate that the Sushila and Rayleigh distributions mixture provides a better fit for bladder cancer remission times compared to other existing distributions.

Parameters Estimation of Bayesian and E-Bayesian Methods on the Generalized Order Statistics under Exponential Family

Parmil Kumar — 2024-06-29

In this paper, we have estimated the parameter of Generalized Order Statistics (GOS) of Exponentiated Distribution Family using Bayesian and E-Bayesian method for computing estimates. To find the estimates, we have employed various loss function viz. Square Error Loss Function (SELF), LINEX and General Entropy Loss Function (GELF). The estimates are derived considering the conjugate prior. Furthermore, the relation among E-Bayesian under different prior distribution of hyperparameters have been established. In the last section, the comparison have been made of derived estimates using Monte Carlo Simulation. To support and validate the obtained result a real data set is analysed.

Using k-means Clustering to Confirm Provincial COVID-19 Cases during the Omicron Epidemic in Thailand

Worrawate Leela-apiradee — 2024-06-29

The Novel Coronavirus 2019 (COVID-19) pandemic has infected and killed millions of people internationally. This work uses k-means clustering and a time series k-means algorithm to present an overview of cases and deaths from COVID-19 in grouped provinces of Thailand before entering the post-pandemic period on 1 July 2022. The study is divided into two parts: the first uses k-means clustering with Euclidean distance measure to analyze confirmed cases and deaths per 100,000 population by province that cumulated from 1 January 2022 to 30 June 2022, during the Omicron (B.1.1.529) outbreak. Based on the elbow method, optimal numeric value for clusters (groups of provinces) is k = 5. The second cluster, consisting of two provinces: Phuket, and Samut Sakhon, is reached the highest cluster mean of the confirmed cases and deaths. We investigate the linear relationship between the confirmed cases (deaths) and 12 different feature variables associated with social, economic, health and environmental factors. Pearson correlation analysis indicates four feature variables correlated positively with confirmed cases and deaths: Gross Regional and Provincial Product (GPP) per capita; number of medical personnel per 100,000 population (pop.); average monthly household income; and number of dengue cases per 100,000 pop. In the second part, k-means clustering with dynamic time warping distance measure is applied to time series data, namely daily confirmed cases per 100,000 people by province gathered during the same time interval as the first part for 181 days, with optimal cluster number being k = 3. The time series of infections attained its apogee in the third cluster, consisting of three provinces: Phuket, Samut Songkhram, and Samut Sakhon. In addition, these findings provide a record of the COVID-19 pandemic in Thailand during the first half of 2022, as illustrated in choropleth maps, for potential governmental use of these provincial groupings for future public health service budget allocation decisions related to the COVID-19 pandemic.

Log-Product-Type Estimator for Estimation of Population Variance Using Auxiliary Information

Prabhakar Mishra — 2024-06-29

This paper proposed a log product type estimator for estimating population variance under simple random sampling without replacement (SRSWOR) using auxiliary information. We have calculated the mean square error (MSE) and bias expressions up to the first order of approximation. To substantiate the result, an empirical study has been performed using three real population data sets. The properties of the estimators also verified through simulation study. The result shows that the performance of the proposed estimator is better than the existing estimators

Explicit Formulas of Average Run Lengths of Moving Average-Range Control Chart

Suganya Phantu — 2024-06-29

Based on the statistical principles, the control chart detects and controls the production process to meet the required quality. This research aimed to present the explicit formula of average run length for moving average based on range control chart for detecting a change of variation. In addition, the efficiency of the change detection of the control chart and R chart at different levels of the parameter change are compared. The criteria used for measuring the efficiency included the average run length for the control process . The research showed that the processes’ results were under normal distribution. The performance of the control chart shows that the chart has lower ARL₁ values than the R chart for all change levels. The adaptation results of the proposed control chart to two sets of actual data corresponded to the research results.

New Product Estimators for Population Mean Under Unequal Probability Sampling with Missing Data: A Case Study on the Number of New COVID-19 Patients

Chugiat Ponkaew — 2024-06-29

The coronavirus pandemic or COVID-19 has killed numerous human lives and the number of COVID-19 patients all over the world including Thailand has drastically increased. Estimating the incidence of COVID-19 can assist in preventing further impacts through policies and planning for the whole nation. Although some information about COVID-19 are missing. If analysis is conducted without dealing with the issue, imprecise estimations may be made from the data. New product estimators along with the variance estimators for estimating population mean have been introduced under unequal probability sampling without replacement with missing data in the study variable under two nonresponse mechansims; missing completely at random and missing at random. Two frameworks are considered; the two-phase and reverse frameworks to find the variance estimators. Simulation studies and an application to COVID-19 patients investigate the performance of the proposed estimators. The results show that the proposed estimators under the missing at random nonresponse mechanism performs the best with the smallest variance compared to other estimators under both frameworks with the estimated mean of new COVID-19 patients equal to 306 cases per week.

Evaluating the Performance of Modified EWMA Control Schemes for Serially Correlated Observations

Rapin Sunthornwat — 2024-06-29

In this study, we propose the explicit formula of the average run length (ARL) for the autoregressive process with a quadratic trend on the modified exponentially weighted moving average (modified EWMA) control chart. The accuracy of the ARL from the explicit formula was compared with the ARL from the numerical integral equation (NIE) method derived by using the quadrature rule. The metrics of comparison were percentage accuracy and computational time. After that, the performance of the modified EWMA control chart is investigated in terms of the average run length, standard deviation of run length (SDRL), and median run length (MRL). In addition, the performance comparison on modified EWMA and the exponentially weighted moving average (EWMA) control charts was presented by using the relative mean index (RMI), the average extra quadratic loss (AEQL), and the performance comparison index (PCI) as the criteria. Furthermore, to determine the ability of the explicit formulas approach, the crude oil WTI price was applied, and it was shown that the modified EWMA control chart performed more significantly than the EWMA control chart under this condition.

Estimation of AQL, LQL and Quality Regions for Group Chain Sampling Plan with Binomial Distribution

Waqar Hafeez — 2024-06-29

Acceptance sampling is a technique for ensuring that both producers and consumers are satisfied with the quality of a product. This paper considers a group chain sampling plan (GChSP) using the binomial distribution. The probability of lot acceptance, is determined by satisfying the producer’s and consumer's risks under the specified design parameters. This paper proves that the proportion of defective decreases when the value of design parameters such as and increase. In this paper for specified values of producer’s and consumer’s risks, four different quality regions are estimated. The findings suggest that for the same values of design parameters the GChSP gives less proportion of defectives than the existing Bayesian group chain sampling plan (BGChSP). Therefore, the GChSP is better equipped for lot inspection in the manufacturing industry, especially those involved with destructive testing of high-quality products.

Wrapped Length Biased Exponential Distribution

Phani Yedlapalli — 2024-06-29

In this article, a new circular distribution called wrapped length biased exponential distribution is introduced. The properties of the new distribution are discussed and explicit expressions are derived for the characteristic function, trigonometric moments, and other statistical measures like resultant length, mean, circular variance, standard deviation, coefficient of skewness, and kurtosis. The maximum likelihood estimation is used to evaluate the model parameter and simulation study is conducted to investigate the performance of the estimator. Finally, an application of the model to a real data set is presented and compared with the fit attained by some other well-known models in the literature.

Estimation of Stress-Strength Reliability of Power Distribution under Type-II Censored Data

Sachin Chaudhary — 2024-06-29

In the statistical literature, there are many lifetime distributions used in reliability analysis, including exponential, normal, gamma, and Weibull distributions. Power distribution is also useful in many scientific contexts, with significant consequences for our understanding of natural and man-made phenomena. This expository paper presents the evaluation of reliability when stress and strength follow power distribution with a common scale and different shape parameters. We obtain maximum likelihood (ML) estimates of stress-strength reliability with their confidence intervals. Furthermore, to compare the performance of various procedures, we apply statistical simulation. Finally, an analysis of a real dataset is given for illustrative purposes.

A New Two-Parameter Extension of Half-Logistic distribution: Properties, Applications and Different Method of Estimations

Majid Hashempour — 2024-06-29

In this paper, we define a new two-parameter lifetime distribution, which is called the new Half-Logistic (NHL) distribution. Theoretical properties of this model, including the hazard function, quantile function, asymptotic, extreme value, moments, conditional moments, mean residual life, mean past lifetime, residual entropy, and order statistics, are derived and studied in detail. The maximum likelihood estimates of parameters are compared with various methods of estimation by conducting a simulation study. Finally, two real data sets illustrate the purposes.