• 한국조사연구학회:학술대회논문집
• /
• 한국조사연구학회 2006년도 춘계학술대회 발표논문집
• /
• pp.131-159
• /
• 2006

This paper consider multinomial group testing which is concerned with classification each of N given units into one of k disjoint categories. In this paper, we propose exact Bayesian, approximate Bayesian, bootstrap methods for estimating individual category proportions using the multinomial group testing model proposed by Bar-Lev et al (2005). By the comparison of Mcan Squre Error (MSE), it is shown that the exact Bayesian method has a bettor efficiency and consistency than maximum likelihood method. We suggest an approximate Bayesian approach using Markov Chain Monte Carlo (MCMC) for posterior computation. We derive exact credible intervals based on the exact Bayesian estimators and present confidence intervals using the bootstrap and MCMC. These intervals arc shown to often have better coverage properties and similar mean lengths to maximum likelihood method already available. Furthermore the proposed models are illustrated using data from a HIV blooding test study throughout California, 2000.

• Journal of the Korean Statistical Society
• /
• 제26권4호
• /
• pp.453-465
• /
• 1997

The family of power-divergence statistics conditional on margins is considered for testing homogeneity of .tau. multinomial populations with equal sample size and the exact Bahadur slope is obtained. It is shown that the likelihood ratio test conditional on margins is the most Bahadur efficient among the family of power-divergence statistics.

• Communications for Statistical Applications and Methods
• /
• 제15권3호
• /
• pp.403-409
• /
• 2008

In this paper we show that the peak of the type I error rate of the Oochran-Armitage trend test could be greater than the nominal level when $2\;{\times}\;C$ contingency tables obtained from two multinomial distributions are extremely unbalanced. This result justifies the use of the exact Cochran-Armitage trend test in extremely unbalanced $2\;{\times}\;C$ contingency tables.

• Journal of the Korean Statistical Society
• /
• 제33권3호
• /
• pp.313-321
• /
• 2004

We consider the problem of testing cell probabilities in sparse multinomial data. Aerts et al. (2000) presented T=${{\Sigma}_{i=1}}^{k}{[{p_i}^{*}-E{(p_{i}}^{*})]^2$ as a test statistic with the local least square polynomial estimator ${{p}_{i}}^{*}$, and derived its asymptotic distribution. The local least square estimator may produce negative estimates for cell probabilities. The local maximum likelihood polynomial estimator ${{\hat{p}}_{i}}$, however, guarantees positive estimates for cell probabilities and has the same asymptotic performance as the local least square estimator (Baek and Park, 2003). When there are cell probabilities with relatively much different sizes, the same contribution of the difference between the estimator and the hypothetical probability at each cell in their test statistic would not be proper to measure the total goodness-of-fit. We consider a Pearson type of goodness-of-fit test statistic, $T_1={{\Sigma}_{i=1}}^{k}{[{p_i}^{*}-E{(p_{i}}^{*})]^2/p_{i}$ instead, and show it follows an asymptotic normal distribution. Also we investigate the asymptotic normality of $T_2={{\Sigma}_{i=1}}^{k}{[{p_i}^{*}-E{(p_{i}}^{*})]^2/p_{i}$ where the minimum expected cell frequency is very small.

• 응용통계연구
• /
• 제26권6호
• /
• pp.889-902
• /
• 2013

지금까지 제안되어 있는 다항 로짓 회귀모형의 적합도 검정 방법들에 대하여 저자들이 제안한 방법들이 타당한지를 확인하고자 본 연구를 진행하였다. 여러 검정 통계량들 중 그룹화 전략을 이용한 통계량들 (Fagerland 등, 2008; Bull, 1994; Pigeon과 Heyse, 1999)을 선정하였고, 이러한 통계량의 기반이 되는 피어슨 ${\chi}^2$ 통계량 또한 같이 비교하였다. 제안된 분포가 모의실험의 상황 하에 얻어지는 귀무분포와 유사한지, 그리고 부적절한 모형의 판별을 적절히 수행하는지에 대하여 확인하였으며, 실제 자료에 세 가지 방법을 적용한 결과를 비교, 평가하였다.

• Communications for Statistical Applications and Methods
• /
• 제17권4호
• /
• pp.527-540
• /
• 2010

규칙기반 분류분석(rule-based classification analysis)은 직관적인 이해가 쉽고 알고리즘이 복잡하지 않아 최근 대용량 데이터마이닝에 많이 이용되는 기법이다. 하지만 현재의 규칙기반 분석은 여러 개의 규칙들을 찾은후 이 규칙들을 단순히 다수결이나 또는 중요도의 가중 합으로서 새로운 데이터를 분류한다. 본 연구에서는 다항분포를 이용한 이항데이터의 분류분석 기법을 규칙 조합방법에 응용하고자한다. 다향분포의 추정을 위해서는 변형된 반복 비율 적합(iterative proportional fitting; IPF) 알고리즘을 이용하여 최대 엔트로피 분포(entropy distribution)를 찾는다. 시뮬레이션 실험 결과 이 방법은 두 집단의 데이터가 서로 유사한 경우 어느 정도 의미 있는 분류 결과를 보여주였다.

• Journal of the Korean Data and Information Science Society
• /
• 제14권2호
• /
• pp.303-311
• /
• 2003

We consider the problem of testing cell probabilities in sparse multinomial data. Aerts, et al.(2000) presented $T_1=\sum\limits_{i=1}^k(\hat{p}_i-p_i)^2$ as a test statistic with the local polynomial estimator $(\hat{p}_i$, and showed its asymptotic distribution. When there are cell probabilities with relatively much different sizes, the same contribution of the difference between the estimator and the hypothetical probability at each cell in their test statistic would not be proper to measure the total goodness-of-fit. We consider a Pearson type of goodness-of-fit test statistic, $T=\sum\limits_{i=1}^k(\hat{p}_i-p_i)^2/p_i$ instead, and show it follows an asymptotic normal distribution.

• 대한교통학회지
• /
• 제1권1호
• /
• pp.43-47
• /
• 1983

Within the conventional transportation planning process, "trip distribution" has a significant role to play. The most widely applied trip distribution model is the gravity model, for which Wilson provided the theoretical basis in 1967. The concept of the gravity model, however, still remains ambiguous if we analyze the "trip distribution" with a disaggregate data set. Thus, this paper hypothesizes that the gravity technique is still valid even with the disaggregate data set, by proving that the estimated coefficients of the gravity model, which is derived under the principle of entropy maximization, are identical with those of the multinomial logit model, which is derived under the principle of individual utility maximization.tility maximization.

• Communications for Statistical Applications and Methods
• /
• 제21권2호
• /
• pp.135-145
• /
• 2014

The present study surveys Bayesian modeling structure for inferences about transition probabilities of Markov chain. The motivation of the study came from the data that shows transitional behaviors of emotionally disturbed children undergoing residential treatment program. Dirichlet distribution was used as prior for the multinomial distribution. The analysis with real data was implemented in WinBUGS programming environment. The performance of the model was compared to that of alternative approaches.

• 환경영향평가
• /
• 제19권2호
• /
• pp.117-125
• /
• 2010

Korea's researchers have recently studied the prediction of forest change, but they have not considered landuse/cover change compared to distribution of forest vegetation. The purpose of our study is to predict forest vegetation based on landuse/cover change on the Korean Peninsula in the 2090's. The methods of this study were Multi-layer perceptrom neural network for Landuse/cover (water, urban, barren, wetland, grass, forest, agriculture) change and Multinomial Logit Model for distribution prediction for forest vegetation (Pinus densiflora, Quercus Spp., Alpine Plants, Evergreen Broad-Leaved Plants). The classification accuracy of landuse/cover change on the Korean Peninsula was 71.3%. Urban areas expanded with large cities as the central, but forest and agriculture area contracted by 6%. The distribution model of forest vegetation has 63.6% prediction accuracy. Pinus densiflora and evergreen broad-leaved plants increased but Quercus Spp. and alpine plants decreased from the model. Finally, the results of forest vegetation based on landuse/cover change increased Pinus densiflora to 38.9% and evergreen broad-leaved plants to 70% when it is compared to the current climate. But Quercus Spp. decreased 10.2% and alpine plants disappeared almost completely for most of the Korean Peninsula. These results were difficult to make a distinction between the increase of Pinus densiflora and the decrease of Quercus Spp. because of they both inhabit a similar environment on the Korean Peninsula.