• Journal of the Korean Data and Information Science Society
• /
• v.23 no.1
• /
• pp.99-111
• /
• 2012

The log-density ratio of the conditional densities of the predictors given the response variable provides useful information for variable selection in the logistic regression model. In this paper, we consider the predictors that are needed and how they should be included in the model. If the conditional distributions are skewed, the distributions can be considered as gamma distributions. Under this assumption, linear and log terms are generally included in the model. The log-odds graph is a very useful graphical tool in this study. A graphical study is presented which shows that if the conditional distributions of x|y for the two groups overlap significantly, we need both the linear and quadratic terms. On the contrary, if they are well separated, only the linear or log term is needed in the model.

• Journal of the Korean Geographical Society
• /
• v.44 no.3
• /
• pp.395-409
• /
• 2009

The objective of this paper is to illustrate that kriging can provide an effective framework both for integrating remote sensing data and for uncertainty modeling through a case study of sediment grain size mapping with remote sensing data. Landsat TM data which show reasonable relationships with grain size values are used as secondary information for sediment grain size mapping near the eastern part of Anmyeondo and Cheonsuman bay. The case study results showed that uncertainty attached to prediction at unsampled locations was significantly reduced by integrating remote sensing data through the analysis of conditional variance from conditional cumulative distribution functions. It is expected that the kriging-based approach presented in this paper would be efficient integration and analysis methodologies for any environmental thematic mapping using secondary information as well as sediment grain size mapping.

• The Korean Journal of Applied Statistics
• /
• v.11 no.2
• /
• pp.247-255
• /
• 1998

In this paper, a new subset selection problem in the Poisson model is considered under the normal predictors. It turns out that the subset model has bigger valiance than that of the Poisson model with random predictors and this has been used to derive new subset selection method similar to Mallows'$C_p$.

• The Korean Journal of Applied Statistics
• /
• v.11 no.2
• /
• pp.257-268
• /
• 1998

Many epidemiological studies investigate whether an association exists between a binary risk factor X and a binary response variable Y. They analyse whether an observed association between X and Y persists when the level of another factor Z that might influence the association is controlled. This involves testing conditional independence of X and Y controlling for Z. The Mantel-Haenszel test is most widely used to test conditional independence for sparse tables. But if the association between X and Y varies along the levels of Z, Mantel-Haenszel test has a low power problem. In this study, we propose an alternative test procedure which overcomes the low power problem in that case. We find out the null distribution of the alternative test statistic and compare its performance with the Mantel-Haenszel test by simulation.

• Tunnel and Underground Space
• /
• v.15 no.2 s.55
• /
• pp.129-136
• /
• 2005

This paper describes the implementation of rock mass rating evaluation based on genetic algorithm(GA) and conditional simulation technique to estimate RMR in the area without sufficient borehole data RMR were estimated by GA and conditional simulation technique with reflecting distribution feature and spatial correlation. And RMR determined by GA were compared with the results from kriging. Through the analysis of the results from 30 simulations, the uncertainty of estimation could be quantified.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.1
• /
• pp.177-185
• /
• 2014

Under bivariate normal distribution assumptions, the interaction and quadratic terms are needed in the logistic regression model with two predictors. However, depending on the correlation coefficient and the variances of two conditional distributions, the interaction and quadratic terms may not be necessary. Although the need for these terms can be determined by comparing the two scatter plots, it is not as useful for interaction terms. We explore the structure and usefulness of the 3-D residual plot as a tool for dealing with interaction in logistic regression models. If predictors have an interaction effect, a 3-D residual plot can show the effect. This is illustrated by simulated and real data.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2007.05a
• /
• pp.1783-1787
• /
• 2007

In this paper we applied radar rainfall for assessment that radar can be used for flood forecasting. The radar data observed at Imjin-River radar site was adjusted using conditional merging method to estimate simulated runoff in Anseon-cheon basin. Also we use two dimensional physical and grid based model call $Vflo^{TM}$. As a result we could find simulated hydrologic curve shows good fitting with observed hydrologic curve even parameters of the model were not calibrated. If we calibrate the parameters, we can expect better hydrologic curve. And radar rainfall can be used for water resources fields and flood forecasting in Korea.

• Journal of the Korean Geotechnical Society
• /
• v.28 no.1
• /
• pp.17-27
• /
• 2012

Since geologic data are often sampled at sparse locations, it is important not only to predict attribute values at unsampled locations but also to assess the uncertainty attached to the prediction. In this study the assessment of the local uncertainty of prediction for the thickness of the consolidation layer was performed by using the indicator approach. A conditional cumulative distribution function (ccdf) was first modeled, and then E-type estimates and the conditional variance were computed for the spatial distribution of the thickness of the consolidation layer. These results could be used to estimate the spatial distribution of secondary compression and to assess the local uncertainty of secondary compression for Songdo New City.

• Journal of the Korean Geotechnical Society
• /
• v.28 no.11
• /
• pp.5-16
• /
• 2012

Since geologic data are often sampled at sparse locations, it is important not only to predict attribute values at unsampled locations, but also to assess the uncertainty attached to the prediction. In this paper, the assessment of the local uncertainty of prediction for the depth to weathered soil was performed by using the indicator kriging. A conditional cumulative distribution function (ccdf) was first modeled, and then E-type estimate was computed for the spatial distribution of the depth to the weathered soil. Also, optimal estimate of spatial distribution for the depth to weathered soil was determined by using ccdf and loss function. The design procedure and method considering the minimum expected loss presented in this paper can be used in the decision-making process for geotechnical engineering design.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09b
• /
• pp.53-56
• /
• 2003

유사도 혹은 자료간의 거리 개념은 많은 기계학습 알고리즘에서 사용되고 있는 중요한 측정개념이다 하지만 입력되는 자료의 속성들중 순서가 정의되지 않은 범주형 속성이 포함되어 있는 경우, 자료간의 유사도나 거리 측정에 어려움이 따른다. 비거리 기반의 알고리즘들의 경우-C4.5, CART-거리의 측정없이 작동할 수 있지만, 거리기반의 알고리즘들의 경우 범주형 속성의 거리 정보 결여로 효과적으로 적용될 수 없는 문제점을 갖고 있다. 본 논문에서는 이러한 범주형 자료들간 거리 측정을 자료 집합의 특성을 충분히 고려한 방법을 제안한다. 이를 위해 자료 집합의 선험적인 정보를 필요로 한다. 이런 선험적 정보인 조건부 확률을 기반으로한 거리 측정방법을 제시하고 오류 피드백을 통해서 속성 간 거리 측정을 최적화 하려고 노력한다. 주어진 자료 집합에 대해 서로 다른 두 범주형 값이 목적 속성에 대해서 유사한 분포를 보인다면 이들 값들은 비교적 가까운 거리로 결정한다 이렇게 결정된 거리를 기반으로 학습 단계를 진행하며 이때 발생한 오류들에 대해 피드백 작업을 진행한다. UCI Machine Learning Repository의 자료들을 이용한 실험 결과를 통해 제안한 거리 측정 방법의 우수한 성능을 확인하였다.