Earticle

최근 환자 의료 데이터를 바탕으로 환자 맞춤형 정밀의료를 수행해 더 나은 치료를 수행하려는 시도가 이어지고 있 다. 그중에서도, 유전자 분석의 비용과 시간이 감소함에 따라 유전체 기반 맞춤 치료 연구가 활발하다. 그러나 암세 포마다 수 천개 존재하는 유전자 변이들의 효과를 직접 파악하는 것은 어렵기 때문에 기존 기법들은 환자 데이터로 부터 통계적 연관성에 기반해 분석을 수행한다. 본 연구는 더 나아가 특정 항암치료의 효과를 높이는 유전자 변이를 탐색하고 중요도에 따라 순위를 매기는 기법을 제안한다. 이를 위해서 개별 유전자가 환자 치료에 미치는 영향을 정 량화하기 위해 비례위험모형 기반의 치료함수를 설계하고, 강화학습 기반의 특징 선택 기법을 도입해 수많은 후보로 부터 유효한 유전자를 추출한다. 분석 결과 제안한 알고리즘은 의학적 지식 없이도 유효한 유전자 변이를 탐색할 수 있음을 보인다.

Recently, there has been an increasing number of attempts to perform the better treatment by performing patient-specific precision medicine. Genome-based customized treatments, one of the representative precision medicines, are receiving much attention as the cost and time of genetic analysis are dramatically decreased. However, since it is challenging to directly grasp the effects of thousands of gene variants present in cancer cell, existing methods perform analysis based on statistical associations from patient data. This study proposes a novel technique for finding gene variants which increase the effectiveness of treatment and ranking them according to their importance. In order to quantify the effect of individual genes on patient treatment, we design the Treatment Effect function based on a proportional hazard model. A feature selection technique based on reinforcement learning is introduced to extract effective genes from numerous gene candidates. We verify that the proposed algorithm can find effective gene variants without any medical knowledge and statistical analysis techniques.

본 연구의 목적은 한국사회적기업진흥원에 공시되 자율경영공시 자료를 이용한 사회적기업 인증진입 가능성을 추정하고 가능성에 미치는 영향요인들을 검증하여 사회적기업 특성을 간접적으로 파악하고자 하였다. 또한 등록제에 대한 논의가 활발하게 진행되고 있는 시점에서 지원제도에 대한 분석으로 사회적기업의 지속가능한 질적 성장과 발전에 기여하고자 한다. 연구방법은 비모수적 분석방법인 카플란마이어 분석으로 생존분석을 실시하였으며 콕스비례위험모형을 이용해 영향요인을 분석하였다. 분석 결과, 먼저 영향요인들이 조직유형과 인증유형, 기업의 규모에 따라 다르게 나타나, 사회적기업 인증제도의 차별된 정책적 차원의 구체적인 유인책과 맞춤형 지원이 필요하다. 둘째, 이사회개최회수의 변수와 사회적기업 인증진입 가능성이 부의 관계로 나타나, 사회적기업이 가지고 있는 의사결정구조 및 구성원에 대한 참여적극성 등을 평가할 수 있는 지표가 필요하다. 이사회개최회수 등과 같은 사회적 성과에 정의 영향을 주는 요인들이, 사회적기업의 이해관계자가 참여하는 의사결정구조에 대한 특성을 반영하면서 성과의 긍정적 효과가 사회적기업 인증진입 가능성과 이행에도 긍정적 효과로 이어질 수 있도록 사회적기업 지원제도의 발전 방향을 설정해야 한다. 셋째, 사회적기업의 자생력 강화를 위해서는 사회적기업 인증진입 가능성에 영향을 주는 요인을 사회적 효과와 경제적 효과를 균형 있게 반영한 지원제도가 필요하고, 매출액이 낮은 집단에 대한 더 지원을 강화하여 취약계층고용 효과를 더 크게 창출할 필요가 있다. 본 연구의 한계점으로는 생존함수의 독립변수 간들의 연계에 대한 검증 부분에 관한 모형 연구 부족과 자료가 기업들이 인증을 받기 위한 제도적 절차상의 자료수집과정 오류 존재에 대하여 한계점으로 제시하였다.

The purpose of this study is to indirectly verify the characteristics of social enterprises by estimating the factors affecting the Possibility of Entering-Certification of Social Enterprises using the Cox Proportional Hazards Model. At a time of active discussion about the registration and certification system, it intends to contribute to the sustainable qualitative growth and development of the social enterprises support system. The research method was survival analysis using a non-parametric Kaplan-Meier. And the factors affecting the possibility of Entering-Certification was analyzed using the Cox Proportional Hazards model. The results were as follows. Firstly, the possibility of Entering-Certification depending on the organization type, certification type, and company size. So the support system requires specific incentives and customized support according to the characteristics of social enterprises. Secondly, the certification system is necessary to reflect social performance, such as decision-making structure and active participation of their members. Thirdly, in order to strengthen the self-reliance of social enterprises, the factors affecting the Possibility of Entering-Certification should be able to reflect the social and economic effects in a balanced way.

This paper considers a condition-based maintenance (i.e. CBM) model for the hydro generating unit in the deregulated power system. As the downtime cost of the generating unit varies according to the time-varying electricity price, the economic dependence among the critical component of the generating unit is dynamic. In this paper, the proposed Dynamic-Economic-Dependence and proportional hazards model based CBM policy considers the dynamic economic dependence among components as well as the reliability and the monitoring information of different critical components. An example for the hydro generating unit is presented to verify the effectiveness of the proposed condition based maintenance policy.

This paper considers a condition-based maintenance (i.e. CBM) model for the hydro generating unit in the deregulated power system. As the generating loss varies according to the time-varying inflows, the economic dependence among the critical component of the generating unit is dynamic. To reduce the maintenance cost effectively, we propose an inflow-dependent control-limit policy instead of the constant control-limit policy for CBM optimization. An example for the hydro generating unit is presented to verify the effectiveness of the proposed policy.

Cox's proportional hazards model (PHM) has been widely applied in the analysis of lifetime data, and it can be characterized by the baseline hazard function and covariates influencing systems' lifetime, where the covariates describe operating environments (e.g. temperature, pressure, humidity). In this article, we consider the constant baseline hazard function and a discrete random variable of a covariate. The estimation procedure is developed in a parametric framework when there are not only complete data but also incomplete one. The Expectation-Maximization (EM) algorithm is employed to handle the incomplete data problem. Simulation results are presented to illustrate the accuracy and some properties of the estimation results.

Proportional hazards model can be used to develop a realistic approach to determine the performance of a system. The proportional hazards model is typically applied for a group of equipments to assess the importance of factors that may influence the reliability of a system. In this paper we considered the interarrival times of a maintained system for the analysis of reliability, maintainability and availability. In order to demonstrate the proposed approach, an example is presented.

Heller and Simonoff(1990) compared several methods of estimating the regression coefficient in a modified proportional hazards model, when the response variable is subject to censoring. We give another method of estimating the parameters in the model which also allows the dependent variable to be censored and the error distribution to be unspecified. The proposed method differs from that of Miller(1976) and that of Buckely and James(1979). We also obtain the variance estimator of the coefficient estimator and compare that with the Buckely-James Variance estimator studied by Hillis(1993).

In this research, we propose a nonparametric test procedure under the proportional hazards assumption for the two-sample problem setting. For this, first of all, we review the functional equation for relative risk between the survival and cumulative functions under the proportional hazards model and the estimate of the relative risk proposed by Park, Hong (2007) with the expressions of the empirical ones for the computational convenience. Then we propose a nonparametric test procedure with the null distribution for the test statistic. For obtaining the null distribution, we apply the large sample approximation theorem and derive the asymptotic normality. Then we compare the performance of our proposed test with that of Begun (1987) with the exponential and Weibull distributions, which are frequently applied in the survival analysis by varying sample sizes. We show that our proposed test procedure can be an excellent alternative to the Begun’s one. Finally we discuss some interesting features of our procedure including re-sampling methods to obtain the null distributions of the test statistics as concluding remarks.

Graphical and numerical methods for checking the assumption of proportional hazards of Cox model for censored survival data are discussed. The strenths and weaknessess of several goodness of fit tests for the propotional hazards for the two-sample problem are evaluated with Monte Carlo simulations, and the tests of Schoenfeld (1980), Andersen (1982), Wei (1984), and Gill and Schumacher (1987) are considered. The goodness of fit methods are illustrated with the survival data of patients who had chronic liver disease and had been treated with the endoscopy injection sclerotheraphy. Two other examples of data known to have nonpropotional hazards are also used in the illustration.

Relapse and factors that affect relapse were analyzed using data on 905 cervical cancer patients by making use of Cox proportional hazards model. The period of this cervical cancer study was conducted from May 1984 to August 2001. Seven categorical and three continuous design variables were measured for each patient. For preliminary statistical analysis, seven design variables were found to be significant at 5% significance level. We reduced the number of variables via backward elimination method to find a survival model to best explain the cervix cancer relapse. The final survival model could be inferred that three factors, disease and depth and size of tumor, among ten design variables had most significant effects on cervix cancer relapse. Relapse rate functions of patients with large tumor were also smoothed.

When we make inferences for survival times for two kinds of treatments, very often we assume the proportionality between hazard functions, which can be described by a constant relative risk. Therefore at the diagnostic stage of data analysis, it becomes important to check the proportionality assumption. We propose a nonparametric test procedure for checking the proportionality assumption between hazard functions using a functional equation. Because of the involvement of censoring distribution function, we consider the large sample case only and obtain the asymptotic normality of the proposed test statistic. Then we discuss the rationale of the use of the functional equation and give some examples. Finally we compare the performances with Andersen's procedure by computing powers through simulations.

We study the asymptotic behavior of the maximum partial likelihood estimator in the Cox proportional hazards model in the presence of nuisance parameters when the entry of patients is staggered. When entry of patients is simultaneous and there is only one regression parameter in the Cox model, the efficient score process of the partial likelihood is martingale and converges weakly to a time-chnaged Brownian motion. Our problem is to get a similar result in the presence of nuisance parameters when entry of patient is staggered.

We propose a simple nonparametric estimator of relative risk in the two sample case of the proportional hazards model for complete data. The asymptotic distribution of this estimator is derived using a functional equation. We obtain the asymptotic normality of the proposed estimator and compare with Begun's estimator by confidence interval through simulations.

In this paper, we investigate some properties of a test procedure, which may be used whether two survival or hazard functions coincide or not under the proportional hazards assumption. Also we discuss that this proposed test can be applied for detecting some specific values of the treatment effect with comparison to the control group. For this purpose, first of all, we review the construction of the test statistic and testing procedure in some detail. Then we comment briefly on the consistency of the test by reasoning that the limiting power tends to unity under the general alternatives. Also we derive the limiting power of test under the scheme of the contiguous alternatives with some appropriate assumptions and the conditions which are known as the Noethter's condition. Then we illustrate the test procedure with a numerical example to test the equality of two underlying distributions. Finally we discuss some interesting aspects related with the proportional hazards model and corresponding tests.

The objective of this study is to investigate the problem of goodness of fit testing based on nonparametric function estimation techniques for the random censorship model. The small and large sample properties of the proposed test, $E_{mn}$, were investigated and it is shown that under the proportional hazard model $E_{mn}$ has higher power compared to the powers of the Kolmogorov -Smirnov, Kuiper, Cramer-von Mises, and analogue of the Cramer-von Mises type test statistic.

This paper presents the asymptotic relative efficiency of the nonparametric estimator relative to the parametric maximum likelihood estimator of the reliability function under the proportional hazards model of random censorship. Also we examine the efficiency loss due to censoring proportions and misson times.

The Cox proportional hazards model is extensively used to model censored survival data. However, relatively little attention has been paid to potential problems arising from the violation of its fundamental assumption; the proportionality assumption that the hazard rates for different values of covariates are proportional over time. When the proportionality assumption does not hold, however, the estimation of the Cox proportional hazards model would be biased. This paper introduces a way to deal with this problem, showing that a non-proportional hazards model can be obtained simply by including in the Cox model the terms for interaction between a function of time and covariates with non-proportional hazards. Then, the model is applied to both simulation data and real data on the adoption of a quality team strategy by business units within a big multinational bank. The analyses confirm that the inattention to non-proportionality of hazards could result in biased estimation and lead to incorrect theoretical interpretation, especially when unobserved heterogeneity is problematic. The paper suggests that proportionality should be carefully checked, and the model that allows hazard ratios to vary over time should be employed.

본 논문에서는 KCTC 훈련 간 생성된 전투원의 생존자료에 대해 통계적 분석모델을 활용하여 생존시간과 생존율에 유의한 영향을 주는 변수들을 찾아내는 생존분석을 실시하였다. 특정 사물의 생존시간과 관련된 다수의 설명변수가 있을 때 위험함수(hazard function)를 기초로 이들의 연관성을 설명하기에 적합한 Cox 비례위험모형을 활용하여, 전투 진행 간 전투원들의 생존시간을 예측하고 생존에 직접적인 영향을 미치는 변수들을 선별하고 검증하였다. 또한, 이분형 종속변수를 갖는 데이터 분석에 이용되는 로지스틱 회귀분석을 활용하여 전투원의 생존시간에 유의한 영향이 있는 변수를 선정하고 검증하는 절차를 동시에 수행하여 분석결과에 대한 신뢰성을 높였다. 이번 연구를 통해 지금까지 전투가 진행됨에 따라 통상적으로 전투원의 생존율(전투력) 추이를 판단하고 훈련에 적용해 왔던 것에서, 전투가 진행되는 특정 시점에서 부대원들의 생존율에 직접적인 영향을 미치는 요인을 선별하고 이를 보완 발전시켜, 각개 전투원의 생존율을 향상시킬 수 있는 객관적인 접근방법을 제시하였다. 이는 군이 보유하고 있는 데이터로부터 새로운 가치와 통찰력을 얻기 위한 최초의 시도라는 점에서 의의가 있다고 할 수 있겠다.

In this paper, survival analysis was conducted to find variables that significantly affect survival time and survival rate by using a statistical analysis model on the survival data of combatants generated during KCTC training. When there are a number of explanatory variables related to the survival time of a specific object, the Cox proportional hazards model, which is suitable to explain their relationship based on the hazard function, is used to predict the survival time of combatants during the battle and directly affect the survival. Influencing variables were selected and verified. In addition, by utilizing logistic regression used for data analysis with binary dependent variables, the procedure for selecting and verifying variables having a significant effect on the survival time of combatants was performed simultaneously to increase the reliability of the analysis results. Through this study, the factors that directly affect the survival rate of troops at a specific point in the battle are selected and analyzed from what has been conventionally applied to training and judging the survival rate(combat power) trend of combatants as battle progresses. By complementing and developing, the survival rate of each combatant was improved, and an objective approach to victory in battle was presented. This is meaningful in that it is the first attempt to obtain new values ??and insights from the data held by the military.