• 대한수학회논문집
• /
• 제34권4호
• /
• pp.1335-1352
• /
• 2019

In this article, a new family of lifetime distributions by adding two additional parameters is introduced. The new family is called, the logarithmic Kumaraswamy family of distributions. For the proposed family, explicit expressions for some mathematical properties are derived. Maximum likelihood estimates of the model parameters are also obtained. This method is applied to develop a new lifetime model, called the logarithmic Kumaraswamy Weibull distribution. The proposed model is very flexible and capable of modeling data with increasing, decreasing, unimodal or modified unimodal shaped hazard rates. To access the behavior of the model parameters, a simulation study has been carried out. Finally, the potentiality of the new method is proved via analyzing two real data sets.

• International Journal of Reliability and Applications
• /
• 제14권2호
• /
• pp.97-105
• /
• 2013

A stress-strength model is formulated for a multi-component system consisting of k identical components. The k components of the system with random strengths ($X_1$, $X_2$, ${\ldots}$, $X_k$) are subjected to one of the r random stresses ($X_{k+1}$, $X_{k+2}$, ${\ldots}$, $X_{k+r}$). The estimation of system reliability based on maximum likelihood estimates (MLEs) and Bayes estimators in k component system are obtained when the system is either parallel or series with the assumption that strengths and stresses follow exponential distribution. A simulation study is conducted to compare MLE and Bayes estimator through the mean squared errors of the estimators.

• Communications for Statistical Applications and Methods
• /
• 제25권6호
• /
• pp.633-645
• /
• 2018

Gaussian error distributions are a common choice in traditional regression models for the maximum likelihood (ML) method. However, this distributional assumption is often suspicious especially when the error distribution is skewed or has heavy tails. In both cases, the ML method under normality could break down or lose efficiency. In this paper, we consider the log-concave and Gaussian scale mixture distributions for error distributions. For the log-concave errors, we propose to use a smoothed maximum likelihood estimator for stable and faster computation. Based on this, we perform comparative simulation studies to see the performance of coefficient estimates under normal, Gaussian scale mixture, and log-concave errors. In addition, we also consider real data analysis using Stack loss plant data and Korean labor and income panel data.

• Journal of information and communication convergence engineering
• /
• 제19권2호
• /
• pp.102-107
• /
• 2021

In this paper, we propose a new three-dimensional (3D) photon-counting integral imaging reconstruction method using a merging reconstruction process and maximum likelihood estimation (MLE). The conventional 3D photon-counting reconstruction method extracts photons from elemental images using a Poisson random process and estimates the scene using statistical methods such as MLE. However, it can reduce the photon levels because of an average overlapping calculation. Thus, it may not visualize 3D objects in severely low light environments. In addition, it may not generate high-quality reconstructed 3D images when the number of elemental images is insufficient. To solve these problems, we propose a new 3D photon-counting merging reconstruction method using MLE. It can visualize 3D objects without photon-level loss through a proposed overlapping calculation during the reconstruction process. We confirmed the image quality of our proposed method by performing optical experiments.

• East Asian Economic Review
• /
• 제26권2호
• /
• pp.143-169
• /
• 2022

This study investigates the effect of economic factors on immigration using the gravity model of immigration. Cross-sectional regression and panel data analyses are conducted from 2000 to 2019 using the OECD International Migration Database, which consists of 36 destination countries and 201 countries of origin. The Poisson pseudo-maximum-likelihood method, which can effectively correct potential biased estimates caused by zeros in the immigration data, is used for estimation. The results indicate that the economic factors strengthened after the global financial crisis. Additionally, this effect varies depending on the type of immigration (the income level of origin country). The gravity model applied to immigration performs reasonably well, but it is necessary to consider the country-specific and time-varying characteristics.

• 전자공학회논문지B
• /
• 제31B권12호
• /
• pp.108-115
• /
• 1994

Most pattern classifiers have been designed based on the ML (Maximum Likelihood) training algorithm which is simple and relatively powerful. The ML training is an efficient algorithm to individually estimate the model parameters of each class under the assumption that all class models in a classifier are statistically independent. That assumption, however, is not valid in many real situations, which degrades the performance of the classifier. In this paper, we propose a minimum-error-rate training algorithm based on the MAP (Maximum a Posteriori) approach. The algorithm regards the normalized outputs of the classifier as estimates of the a posteriori probability, and tries to maximize those estimates. According to Bayes decision theory, the proposed algorithm satisfies the condition of minimum-error-rate classificatin. We apply this algorithm to NPM (Neural Prediction Model) for speech recognition, and derive new disrminative training algorithms. Experimental results on ten Korean digits recognition have shown the reduction of 37.5% of the number of recognition errors.

• Communications for Statistical Applications and Methods
• /
• 제15권2호
• /
• pp.161-171
• /
• 2008

This study discusses Johnson's $S_U$-normal distribution capturing a wide range of non-normality in various regression models. We provide the likelihood inference using Johnson's $S_U$-normal distribution, and propose a likelihood ratio (LR) test for normality. We also apply the $S_U$-normal distribution to the binary and censored regression models. Monte Carlo simulations are used to show that the LR test using the $S_U$-normal distribution can be served as a model specification test for normal error distribution, and that the $S_U$-normal maximum likelihood (ML) estimators tend to yield more reliable marginal effect estimates in the binary and censored model when the error distributions are non-normal.

• Animal Bioscience
• /
• 제37권3호
• /
• pp.419-427
• /
• 2024

Objective: The objective of this study was to obtain (co) variance components and genetic parameter estimates for post-weaning body measurements such as body length (BL), height at withers (HW), and chest girth (HG) recorded at six (SBL, SHW, and SHG), nine (NBL, NHW, and NHG) and twelve (YBL, YHW, and YHG) months of age along with yearling weight (YW) in Nellore sheep maintained at livestock research station, Palamaner, Andhra Pradesh, India and also the association among body measurements with YW was studied. Methods: Data on 2,076 Nellore sheep (descended from 75 sires and 522 dams) recorded between 2007 and 2016 (10 years) were utilized in the study. Lambing year, sex of lamb, season of lambing and parity of dam were included in the model as fixed effects and ewe weight was kept as a covariate. Analyses were conducted with six animal models with different combinations of direct and maternal genetic effects using restricted maximum likelihood procedure. Best model for each trait was determined based on Akaike's information criterion. Results: Moderate estimates of direct heritability were obtained for the studied traits viz., BL (0.02 to 0.24), HW (0.31 to 0.49), and CG (0.08 to 0.35) and their corresponding maternal heritability estimates were in the range of 0.00 to 0.07 (BL), 0.13 to 0.17 (HW), and 0.07 to 0.13 (CG), respectively. Positive direct genetic and phenotypic correlations among the traits and they ranged from 0.07 (YBL-YW) to 0.99 (SBL-SHG, SHG-YW, and NBL-YBL) and 0.01 (SBL-YBL) to 0.99 (NBL-NHG), respectively. Further, the genetic correlations among all the body measurements and YW were positive and ranged from 0.07 (YW and YBL) to 0.99 (YW and SHG). Conclusion: There was a strong association of chest girth at six months with YW. Further, it is indicated that moderate improvement of post-weaning body measurements in Nellore sheep would be possible through selection.

• Journal of Advanced Marine Engineering and Technology
• /
• 제40권2호
• /
• pp.131-137
• /
• 2016

In reliability analysis, it is quite common for the failure of any individual or item to be attributable to more than one cause. Moreover, observed data are often censored. Recently, progressive hybrid censoring schemes have become quite popular in life-testing problems and reliability analysis. However, a limitation of the progressive hybrid censoring scheme is that it cannot be applied when few failures occur before time T. Therefore, generalized progressive hybrid censoring schemes have been introduced. In this article, we derive the likelihood inference of the unknown parameters under the assumptions that the lifetime distributions of different causes are independent and exponentially distributed. We obtain the maximum likelihood estimators of the unknown parameters in exact forms. Asymptotic confidence intervals are also proposed. Bayes estimates and credible intervals of the unknown parameters are obtained under the assumption of gamma priors on the unknown parameters. Different methods are compared using Monte Carlo simulations. One real data set is analyzed for illustrative purposes.

• 대한간호학회지
• /
• 제29권5호
• /
• pp.1030-1041
• /
• 1999

This study examined those problems noticed under the Asymptotically Generalized Least Squares estimator in evaluating a structural model of physical health. The problems were highly correlated parameter estimates and high standard errors of some parameter estimates. Separate analyses of the endogenous part of the model and of the metric of a latent factor revealed a highly skewed and kurtotic measurement indicator as the focal point of the manifested problems. Since the sample sizes are far below that needed to produce adequate AGLS estimates in the given modeling conditions, the adequacy of the Maximum Likelihood estimator is further examined with the robust statistics and the bootstrap method. These methods demonstrated that the ML methods were unbiased and statistical decisions based upon the ML standard errors remained almost the same. Suggestions are made for future studies adopting structural equation modeling technique in terms of selecting of a reference indicator and adopting those statistics corrected for nonormality.