• The Korean Journal of Applied Statistics
• /
• v.33 no.3
• /
• pp.335-345
• /
• 2020

This paper deals with the problem of estimating the log-transformed linear regression model to fit actual battle data from the Ardennes Campaign of World War II into the Lanchester model. The problem of determining a global solution for parameters and multicollinearity problems are identified and modified by examining the results of previous studies on data. The least squares method requires attention because a local solution can be found rather than a global solution if considering a specific constraint or a limited candidate group. The method of exploring this multicollinearity problem can be confirmed by a statistic known as a variance inflation factor. Therefore, the Lanchester model is simplified to avoid these problems, and the combat power attrition rate model was proposed which is statistically significant and easy to explain. When fitting the model, the dependence problem between the data has occurred due to autocorrelation. Matters that might be underestimated or overestimated were resolved by the Cochrane-Orcutt method as well as guaranteeing independence and normality.

• Journal of the Korean Statistical Society
• /
• v.8 no.2
• /
• pp.109-116
• /
• 1979

두 개의 변수의 대체효과(substitution effect)를 연구하기 위하여 수요 또는 공급의 모형을 만들었을 경우 이에 관련된 변수들의 이름이 중요시 된다. 실제 관측 자료를 사용하였을 경우 흔히 일어나는 다공선성(multicollinearity) 문제를 다루기 위한 대안으로써 선형회귀선을 예로 들어 능형회귀기법(ridge regression technique)과 요인분석기법(factor analytic technique)을 소개하였으며 이에서 얻어지는 계수(coefficient)를 OLS 추정치로 설명하기 위하여 원래의 자료를 변환하였다. 실지 수요와 공급의 모형이 비선형일 경우 일반적으로 능형회귀나 요인분석을 쓰지 못한다는 점을 감안, 이러한 방법을 자료의 변환방법으로 설명함으로써 비선형모형에서도 다공선성문제를 위하여 능형회귀분석법이나 요인분석기법을 사용할 수 있도록 하였다.

• Journal of applied mathematics & informatics
• /
• v.6 no.2
• /
• pp.599-609
• /
• 1999

Neural networks are explored as an alternative to a regres-sion model for prediction of the number of daily household vehicular trips. This study focuses on contrasting a neural network model with a regression model in term of variable selection as well as the appli-cation of these models for prediction of extreme observations, The differences in the models regarding data transformation variable selec-tion and multicollinearity are considered. The results indicate that the neural network model is a viable alternative to the regression model for addressing both messy data problems and limitation in variable structure specification.

• 대한예방의학회:학술대회논문집
• /
• 1993.10a
• /
• pp.41-42
• /
• 1993

• Journal of the Korean Data and Information Science Society
• /
• v.11 no.1
• /
• pp.73-81
• /
• 2000

The variance inflation factor can be expressed by the square of the ratio of t-statistics associated with slopes of partial regression and partial residual plots. Disagreement of two sides in the interpretation can be occurred, and we analyze it with some illustrations.

• Proceedings of the Korea Contents Association Conference
• /
• 2014.11a
• /
• pp.461-462
• /
• 2014

기존 전문가 델파이 평가를 대체하는 특허기반 기술수준 평가모형들의 독립변수로 활용되는 특허활동도, 특허집중도, 특허시장력, 특허경쟁력 및 특허영향력의 다중공선성이 존재하여 이를 제거함으로써 보다 신뢰성이 높은 기술수준 평가모형을 실증하여 제안하고자 한다.

• Journal of Korean Society of Transportation
• /
• v.30 no.1
• /
• pp.113-123
• /
• 2012

This study introduces a new type of a modal split model that use zonal data instead of cost data as independent variables. It has been indicated that the ones using cost data have deficiencies in the multicollinearity of travel time and cost variables and unpredictability of independent variables. The zonal data employed in this study include (1) socioeconomic data, (2) land use data and (3) transportation system data. The test results showed that the proposed modal split model using zonal data performs better than the other does.

• The Korean Journal of Applied Statistics
• /
• v.27 no.4
• /
• pp.577-588
• /
• 2014

The collection and storage of a large volumes of data are becoming easier; however, the number of variables is sometimes more than the number of samples(objects). We now face the problem of dependency among variables(such as multicollinearity) due to the increased number of variables. We cannot apply various statistical methods without satisfying independency assumption. In order to overcome such a drawback we consider a categorizing (or discretizing) observations. We have a data set of nancial indices from banks in Korea that contain 78 variables from 16 banks. Genetic sequence data is also a good example of a large data and there have been numerous statistical methods to handle it. We discover lots of useful bank information after we transform bank data into categorical data that resembles genetic sequence data and apply bioinformatic techniques.

• IE interfaces
• /
• v.25 no.3
• /
• pp.347-356
• /
• 2012

The acquisition cost of defense weapon system has been continuously increasing because of art-of-technology of it. This phenomenon requires efficiency and transparency in the weapon system acquisition process through cost estimation. Therefore cost estimation is very important to the government acquisition programs to support decisions about funding and to evaluate resource requirement as a key decision point. The Commercial parametric cost estimating models have been using extensively to obtain appropriate cost estimates in early acquisition phase. These models have many restrictions to ensure the cost estimating result in Korean defense environment because they are developed based on foreign R&D data. Also estimation results are different from Korean defense industry accounting system. So, some studies have been tried to develop a CER (Cost Estimation Relationship) based on the Korean historical data. However, there are some restrictions to improve the predictability and ensure the stability of the developed singular CERs which consider the following data characteristics individually. The the abnormal conditions of data that is multicollinearity, outlier and heteroscedasticity under rack of the number of observations. In this paper, a CER's Linear Combining Model is proposed to overcome those limitations which guarantee more accurate estimation (25.42% higher precision) than other singular CERs. At least, this study is meaningful as a first attempt to improve the predictability of CER with insufficient data. The methodology suggested in this study will be useful to develop a complex Korean version cost estimating model development in future.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.1
• /
• pp.1-10
• /
• 2017

Longitudinal data of Korean child academic achievement have been used to find the significant exploratory variables under the assumption of independent repeated measured data. Using the exploratory variables in previous research works, we analyze the linear mixed model incorporating the fixed and random effects for child academic achievement to detect the significant exploratory variables. Korea welfare panel study data observed three times between 2006 and 2012 by additional survey for children. The child academic achievement is evaluated by the sum of academic achievements of Korean, English and Mathematics. We also investigate the multicollinearity and the missing mechanism and select some popular correlation matrices to analyze the linear mixed model.