• Proceedings of the Korea Contents Association Conference
• /
• 2013.05a
• /
• pp.171-172
• /
• 2013

우울은 군대 내 발생되는 극단적인 사고 중 하나인 자살의 주요 원인으로 제시되어 왔다. 본 연구는 군인들의 우울, 불안 및 자아존중감의 수준을 파악하고, 우울의 영향요인을 탐색하고 이들을 예측하는데 주로 사용해 왔던 다중회귀분석 방법과 효과적인 의사결정방법으로 알려진 회귀나무모형의 효과성을 비교해보고자 하였다. 방법: 횡단적 조사연구이며, 우울측정에는 CES-D, 불안측정은 SAI, 자아존중감은 Rosenberg(1965)의 도구를 사용하였다. 연구대상자는 강원도 전방 부대 근무 중인 군인이며, 534부가 회수되었다. SPSS/WIN 18.0을 이용하여 위계적 다중회귀분석과 회귀나무모형을 실시하였다. 결과: 대상자들의 우울, 불안 및 자아존중감의 정도는 각각 $10.7({\pm}9.8)$, $38.5({\pm}10.2)$과 $31.7({\pm}5.2)$이었다. 대상자의 23.6%(126명)가 경한 우울을 나타내었다. 다중회귀분석에 의한 우울 영향요인은 불안, 자아존중감과 복무기간이었으며, 우울에 대하여 62.0%의 설명력을 가지고 있었다. 또한 회귀나무모형에서는 높은 불안과 불안이 다소 낮더라도 전역 후 진로가 불확실한 집단이 우울 위험군일 것으로 예측되었다. 결론: 본 연구 대상자들의 우울의 주요 영향요인은 불안으로 나타났다. 군대 내에서 적용할 수 있는 불안 조절 방법 개발이 필요할 것으로 보인다. 또한 일부 요인에서 차이가 있어, 반복 연구가 필요하지만, 주요 변인인 불안을 예측했다는 점에서 보면 다중회귀분석과 회귀나무모형은 군인들의 우울을 예측에 유용한 방법으로 보인다.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.6
• /
• pp.1253-1262
• /
• 2014

Regression tree is a tree-structured solution in which a simple regression model is fitted to the data in each node made by recursive partitioning of predictor space. There have been many efforts to apply tree algorithms to various regression problems like logistic regression and quantile regression. Recently, algorithms have been expanded to the panel data analysis such as RE-EM algorithm by Sela and Simonoff (2012), and extension of GUIDE by Loh and Zheng (2013). The algorithms are briefly introduced and prediction accuracy of three methods are compared in this paper. In general, RE-EM shows good prediction accuracy with least MSE's in the simulation study. A RE-EM tree fitted to business survey index (BSI) panel data shows that sales BSI is the main factor which affects business entrepreneurs' economic sentiment. The economic sentiment BSI of non-manufacturing industries is higher than that of manufacturing ones among the relatively high sales group.

• The Korean Journal of Applied Statistics
• /
• v.33 no.6
• /
• pp.753-762
• /
• 2020

With an increasing interest in integrating artificial intelligence (AI) into interview processes, the Republic of Korea (ROK) army is trying to lead and analyze AI-powered interview platform. This study is to analyze the AI interview data using a unified non-crossing multiple quantile tree (UNQRT) model. Compared to the UNQRT, the existing models, such as quantile regression and quantile regression tree model (QRT), are inadequate for the analysis of AI interview data. Specially, the linearity assumption of the quantile regression is overly strong for the aforementioned application. While the QRT model seems to be applicable by relaxing the linearity assumption, it suffers from crossing problems among estimated quantile functions and leads to an uninterpretable model. The UNQRT circumvents the crossing problem of quantile functions by simultaneously estimating multiple quantile functions with a non-crossing constraint and is robust from extreme quantiles. Furthermore, the single tree construction from the UNQRT leads to an interpretable model compared to the QRT model. In this study, by using the UNQRT, we explored the relationship between the results of the Army AI interview system and the existing personnel data to derive meaningful results.

• The Korean Journal of Applied Statistics
• /
• v.29 no.7
• /
• pp.1361-1371
• /
• 2016

Quantile regression provides a variety of useful statistical information to examine how covariates influence the conditional quantile functions of a response variable. However, traditional quantile regression (which assume a linear model) is not appropriate when the relationship between the response and the covariates is a nonlinear. It is also necessary to conduct variable selection for high dimensional data or strongly correlated covariates. In this paper, we propose a penalized quantile regression tree model. The split rule of the proposed method is based on residual analysis, which has a negligible bias to select a split variable and reasonable computational cost. A simulation study and real data analysis are presented to demonstrate the satisfactory performance and usefulness of the proposed method.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.3
• /
• pp.533-545
• /
• 2017

Quantile regression models provide a variety of useful statistical information by estimating the conditional quantile function of the response variable. However, the traditional linear quantile regression model can lead to the distorted and incorrect results when analysing real data having a nonlinear relationship between the explanatory variables and the response variables. Furthermore, as the complexity of the data increases, it is required to analyse multiple response variables simultaneously with more sophisticated interpretations. For such reasons, we propose a multivariate quantile regression tree model. In this paper, a new split variable selection algorithm is suggested for a multivariate regression tree model. This algorithm can select the split variable more accurately than the previous method without significant selection bias. We investigate the performance of our proposed method with both simulation and real data studies.

• Proceedings of the Korean Statistical Society Conference
• /
• 2003.05a
• /
• pp.103-108
• /
• 2003

데이터마이닝을 위한 대용량 데이터베이스를 축소시키는 방법 중에 속성선택 방법이 많이 사용되고 있다. 본 논문에서는 세 가지 속성선택 방법을 사용하여 조건속성 수를 60%이상 축소시켜 결정나무와 로지스틱 회귀모형에 적용시켜보고 이들의 효율을 비교해 본다. 세 가지 속성선택 방법은 MDI, 정보획득, ReliefF 방법이다. 결정나무 방법은 QUEST, CART, C4.5를 사용하였다. 속성선택 방법들의 분류 정확성은 UCI 데이터베이스에 주어진 Credit 승인 데이터베이스와 German Credit 데이터베이스를 사용하여 10층-교차확인 방법으로 평가하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.4
• /
• pp.777-786
• /
• 2018

This study propose a new method to detect Cochlodinium polykrikoides on satellite images using logistic regression and decision tree. We used spectral profiles(918) extracted from red tide, clear water and turbid water as training data. The 70% of the entire data set was extracted and used for model training, and the classification accuracy of the model was evaluated by using the remaining 30%. As a result of the accuracy evaluation, the logistic regression model showed about 97% classification accuracy, and the decision tree model showed about 86% classification accuracy.

• Journal of Korean Society of Forest Science
• /
• v.112 no.1
• /
• pp.23-31
• /
• 2023

Predicting forest productivity is essential to evaluate sustainable forest management or to enhance forest ecosystem services. Ordinary least squares (OLS) and partial least squares (PLS) regression models were used to develop predictive models for forest productivity (site index) from the site characteristics and soil profile, along with soil physical and chemical properties, of 112 Quercus mongolica stands. The adjusted coefficients of determination (adjusted R²) in the regression models were higher for the site characteristics and soil profile of B horizon (R²=0.32) and of A horizon (R²=0.29) than for the soil physical and chemical properties of B horizon (R²=0.21) and A horizon (R²=0.09). The PLS models (R²=0.20-0.32) were better predictors of site index than the OLS models (R²=0.09-0.31). These results suggest that the regression models for Q. mongolica can be applied to predict the forest productivity, but new variables may need to be developed to enhance the explanatory power of regression models.

• The Korean Journal of Applied Statistics
• /
• v.21 no.4
• /
• pp.639-647
• /
• 2008

This study used the characteristics of the knowledge discovery and data mining algorithms to develop tailored hypertension follow up management model - hypertension care predictive model and hypertension care compliance segmentation model - for hypertension management using the Korea National Health Insurance Corporation database(the insureds’ screening and health care benefit data). This study validated the predictive power of data mining algorithms by comparing the performance of logistic regression, decision tree, and ensemble technique. On the basis of internal and external validation, it was found that the model performance of logistic regression method was the best among the above three techniques on hypertension care predictive model and hypertension care compliance segmentation model was developed by Decision tree analysis. This study produced several factors affecting the outbreak of hypertension using screening. It is considered to be a contributing factor towards the nation’s building of a Hypertension follow up Management System in the near future by bringing forth representative results on the rise and care of hypertension.

• The Journal of the Korea Contents Association
• /
• v.13 no.12
• /
• pp.829-839
• /
• 2013

This study is a descriptive research study with the purpose of predicting and comparing factors of depression affecting residents in a metropolitan city by using logistic regression analysis and decision-making tree analysis. The subjects for the study were 462 residents ($20{\leq}aged{\angle}65$) in a metropolitan city. This study collected data between October 7, 2011 and October 21, 2011 and analyzed them with frequency analysis, percentage, the mean and standard deviation, ${\chi}^2$-test, t-test, logistic regression analysis, roc curve, and a decision-making tree by using SPSS 18.0 program. The common predicting variables of depression in community residents were social dysfunction, perceived physical symptom, and family support. The specialty and sensitivity of logistic regression explained 93.8% and 42.5%. The receiver operating characteristic (roc) curve was used to determine an optimal model. The AUC (area under the curve) was .84. Roc curve was found to be statistically significant (p=<.001). The specialty and sensitivity of decision-making tree analysis were 98.3% and 20.8% respectively. As for the whole classification accuracy, the logistic regression explained 82.0% and the decision making tree analysis explained 80.5%. From the results of this study, it is believed that the sensitivity, the classification accuracy, and the logistics regression analysis as shown in a higher degree may be useful materials to establish a depression prediction model for the community residents.