• 가정과삶의질연구
• /
• 제10권1호통권19호
• /
• pp.131-138
• /
• 1992

Regression analysis has become a standard statistical tool in the behavioral science. Because of its widespread popularity. regression has been often misused. Such is the case when the dependent variable is a qualitative measure rather than a continuous, interval measure. Regression estimates with a qualitative dependent variable does not meet the assumptions underlying regression. It can lead to serious errors in the standard statistical inference. Logit model is recommended as alternatives to the regression model for qualitative dependent variables. Researchers can employ this model to measure the relationship between independent variables and qualitative dependent variables without assuming that logit model was derived from probabilistic choice theory. Coefficients in logit model are typically estimated by the method of Maximum Likelihood Estimation in contrast to ordinary regression model which estimated by the method of Least Squares Estimation. Goodness of fit in logit model is based on the likelihood ratio statistics and the t-statistics is used for testing the null hypothesis.

• Journal of the Korean Statistical Society
• /
• 제30권1호
• /
• pp.41-61
• /
• 2001

The minimum negative exponential disparity estimator(MNEDE), introduced by Lindsay(1994), is an excellenet competitor to the minimum Hellinger distance estimator(Beran 1977) as a robust and yet efficient alternative to the maximum likelihood estimator in parametric models. In this paper we define the negative exponential deviance test(NEDT) as an analog of the likelihood ratio test(LRT), and show that the NEDT is asymptotically equivalent to he LRT at the model and under a sequence of contiguous alternatives. We establish that the asymptotic strong breakdown point for a class of minimum disparity estimators, containing the MNEDE, is at least 1/2 in continuous models. This result leads us to anticipate robustness of the NEDT under data contamination, and we demonstrate it empirically. In fact, in the simulation settings considered here the empirical level of the NEDT show more stability than the Hellinger deviance test(Simpson 1989). The NEDT is illustrated through an example data set. We also define a goodness-of-fit statistic to assess adequacy of a specified parametric model, and establish its asymptotic normality under the null hypothesis.

• 한국음향학회지
• /
• 제27권8호
• /
• pp.443-449
• /
• 2008

이동통신에서 배경잡음이 존재하는 실제 환경에서 음성신호처리의 가장 중요한 이슈중의 하나는 강인한 음성검출기를 설계하는 것이다. 상대적으로 간단하면서도 성능이 우수하여 대표적인 음성검출기로 사용되는 통계적모델기반 기법은 각 주파수 채널별 우도비를 이용하여 음성검출 검출식을 만들어내는 방식이다. 최근, 변별적 가중치 학습 (discriminative weight training)을 이용하여 주파수 체널별 가중치가 인가된 우도비를 이용한 음성검출 결정식을 갖는 음성검출기가 제안 되었으며 상대적으로 우수한 성능을 보였다. 본 연구에서는 기존의 변별적 가중치 학습의 입력벡터에 이전프레임의 결정식을 궤환구조형태를 바탕으로 추가하는 새로운 방식을 제안한다. 제안된 기법은 비정상 (non-staionary) 잡음 환경에서 객관적인 방법을 통해 상호비교 분석되었으며 결론적으로 우수한 성능을 보였다.

• Communications for Statistical Applications and Methods
• /
• 제19권3호
• /
• pp.495-500
• /
• 2012

We consider a Bayesian test of independence in a two-way contingency table that has some zero cells. To do this, we take a three-stage hierarchical Bayesian model under each hypothesis. For prior, we use Dirichlet density to model the marginal cell and each cell probabilities. Our method does not require complicated computation such as a Metropolis-Hastings algorithm to draw samples from each posterior density of parameters. We draw samples using a Gibbs sampler with a grid method. For complicated posterior formulas, we apply the Monte-Carlo integration and the sampling important resampling algorithm. We compare the values of the Bayes factor with the results of a chi-square test and the likelihood ratio test.

• Journal of the Korean Statistical Society
• /
• 제35권4호
• /
• pp.355-376
• /
• 2006

To identify whether the sequence of observations follows a chain dependent process and whether the chain dependent or repeated observations follow stationary process or not, alternative procedures are suggested in this paper. These test procedures are formulated on the basis of logistic regression model under the likelihood ratio test criterion and applied to the daily rainfall occurrence data of Bangladesh for selected stations. These test procedures indicate that the daily rainfall occurrences follow a chain dependent process, and the different types of transition probabilities and overall transition probabilities of Markov chain for the occurrences of rainfall follow a stationary process in the Mymensingh and Rajshahi areas, and non-stationary process in the Chittagong, Faridpur and Satkhira areas.

• 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.

• Journal of the Korean Data and Information Science Society
• /
• 제9권2호
• /
• pp.289-298
• /
• 1998

We investigate the distribution of likelihood ratio test(LRT) of null hypothesis a sample is from single gamma with unknown shape and scale against the alternative hypothesis a sample is from a mixture of two gammas, each with unknown scale and unknown (but equal) scale. To obtain stable maximum likelihood estimates(MLE) of a mixture of two gamma distributions, the EM(Dempster, Laird, and Robin(1977))and Modified Newton(Jensen and Johansen(1991)) algorithms were implemented. Based on EM, we made a simple structure likelihood equation for each parameter and could obtain stable solution by Modified Newton Algorithms. Simulation study was conducted to investigate the distribution of LRT for sample size n = 25, 50, 75, 100, 50, 200, 300, 400, 500 with 2500 replications. To determine the small sample distribution of LRT, I considered the model of a gamma distribution with shape parameter equal to 1 + f(n) and scale parameter equal to 2. The simulation results indicate that the null distribution is essentially invariant to the value of the shape parameter. Modeling of the null distribution indicates that it is well approximated by a gamma distribution with shape parameter equal to the quantity $0.927+1.18/\sqrt{n}$ and scale parameter equal to 2.16.

• Journal of the Korean Statistical Society
• /
• 제19권2호
• /
• pp.171-175
• /
• 1990

A diagnostic test for detecting nonconstant variance in mixed linear models based on the score statistic is derived through the technique of model expansion, and compared to the log likelihood ratio test.

• Journal of the Korean Statistical Society
• /
• 제30권2호
• /
• pp.179-192
• /
• 2001

When a statistical model has a hierarchical structure such as multilayer perceptrons in neural networks or Gaussian mixture density representation, the model includes distribution with unidentifiable parameters when the structure becomes redundant. Since the exact structure is unknown, we need to carry out statistical estimation or learning of parameters in such a model. From the geometrical point of view, distributions specified by unidentifiable parameters become a singular point in the parameter space. The problem has been remarked in many statistical models, and strange behaviors of the likelihood ratio statistics, when the null hypothesis is at a singular point, have been analyzed so far. The present paper studies asymptotic behaviors of the maximum likelihood estimator and the Bayesian predictive estimator, by using a simple cone model, and show that they are completely different from regular statistical models where the Cramer-Rao paradigm holds. At singularities, the Fisher information metric degenerates, implying that the cramer-Rao paradigm does no more hold, and that he classical model selection theory such as AIC and MDL cannot be applied. This paper is a first step to establish a new theory for analyzing the accuracy of estimation or learning at around singularities.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2004년도 Proceedings of ISRS 2004
• /
• pp.314-317
• /
• 2004

The aim of this study was to evaluate the susceptibility from landslides in the Lai Chau region of Vietnam, using Geographic Information System (GIS) and remote sensing data, focusing on the relationship between tectonic fractures and landslides. Landslide locations were identified from an interpretation of aerial photographs and field surveys. Topographic and geological data and satellite images were collected, processed, and constructed into a spatial database using GIS data and image processing techniques, and a scheme of the tectonic fracturing of the crust in the Lai Chau region was established. In this scheme, Lai Chau was identified as a region with low crustal fractures, with the grade of tectonic fracture having a close relationship with landslide occurrence. The factors found to influence landslide occurrence were: topographic slope, topographic aspect, topographic curvature, distance from drainage, lithology, distance from a tectonic fracture and land cover. Landslide prone areas were analyzed and mapped using the landslide occurrence factors employing the probability-likelihood ratio method. The results of the analysis were verified using landslide location data, and these showed a satisfactory agreement between the hazard map and existing landslide location data.