• Journal of the Korean Data and Information Science Society
• /
• v.27 no.5
• /
• pp.1253-1262
• /
• 2016

When there exists a change-point in hazard function, it should be estimated for exact parameter or hazard estimation. In this research, we compare the hazard change-point estimators. Matthews and Farewell (1982) parametric change-point estimator is based on the likelihood and Zhang et al. (2014) nonparametric estimator is based on the Nelson-Aalen cumulative hazard estimator. Simulation study is done for the data from exponential distribution with one hazard change-point. The simulated data generated without censoring and the data with right censoring are considered. As real data applications, the change-point estimates are computed for leukemia data and primary biliary cirrhosis data.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.6
• /
• pp.1225-1238
• /
• 2015

In this research, we fit various survival models and conduct tests and estimation for the hazard change-point with the rectal cancer data. By the log-rank tests, at significance level ${\alpha}=0.10$, survival functions are significantly different according to the uniporter of glucose (GLUT1), clinical stage (cstage) and pathologic stage (ypstage). From the Cox proportional hazard model, the most significant covariates are GLUT1 and ypstage. Assuming that the rectal cancer data follows the exponential distribution, we estimate one hazard change-point using Matthews and Farewell (1982), Henderson (1990) and Loader (1991) methods.

• Communications for Statistical Applications and Methods
• /
• v.5 no.2
• /
• pp.477-489
• /
• 1998

위험률 변화점모형에 대해 변화점의 최우추정을 고려하였다. 추정량의 점근분포 및 붓스트랩 분포의 성질을 알아보고 변화점의 신뢰구간을 제안한다. 변화점의 위치 및 변화점을 전후하여 위험률의 값에 따라 모의실험을 수행하고 포함확률을 조사하였다. 추정량의 점근분포가 매우 복잡하기 때문에 이를 직접 이용한 변화점의 통계적 추론이 매우 어려운 점을 감안할 때 제안된 방법은 바람직한 대안이 될 수 있다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2008.05a
• /
• pp.490-494
• /
• 2008

본 연구에서는 상수도 배수관로의 내 외부적 특성에 따라 개별관로를 정의하는 방법을 연구대상 지역의 배수관로 파손 데이터베이스에 적용하여 비례위험모형을 구축하였다. 연구에 사용된 자료는 연구대상지역의 배수관로의 제원 및 파손시기를 포함하는 관로 파손데이터베이스, 관로매설지역의 급수인구 및 수압범위에 관한 자료를 포함하는 GRID 데이터베이스와 관로매설지역의 토지개발 정도에 관한 자료를 포함한다. 이러한 자료를 이용하여 관로를 순차적 파손경험에 따라 7개의 생존시간군(STG I $\sim$ VII)으로 구분하고 각 생존시간군에 대한 비례위험모형(Model I $\sim$ VII)을 구축하였다. 이러한 모형을 이용하여 관로의 파손횟수가 증가하는 동안 파손에 영향을 미치는 인자의 변화와 그 효과를 파악하였으며, 또한 추정된 공변수의 위험비율을 분석함으로써 관로의 제원 혹은 매설환경, 급수인구 등에 따른 위험률의 상대적인 변화를 분석하였다. 또한 비례 위험모형의 구축과정에서 관로의 파손에 영향을 미치는 공변수의 비례성 가정을 검토하여 시간종속형 공변수를 모형화하였으며, 모형의 이탈잔차(deviance residual)를 분석하여 모형의 적합성을 검토하였다. 본 연구에서 구축된 비례위험모형에 대해 Shoenfeld 잔차를 이용한 스코어 잔차의 변화(score process)를 검토한 결과, Model I 과 Model II 에 대해서는 공변수의 시간종속 효과가 발견되었다. Model I에 대해서는 관로재질과 급수인구의 영향이 시간에 따라 변하며 Model II에서는 급수인구의 영향만이 시간에 따라 변하는 것으로 나타났다. 한편 Model III $\sim$ Model VII 들에 대해서는 공변수의 영향이 시간에 따라 변하지 않는 것으로 나타났다. 각 생존시간군에 대해 관로재질, 토지개발정도, 관로길이 및 급수인구의 변화가 관로의 상대적 누수위험률에 미치는 영향을 상대위험률의 95% 신뢰구간을 고려하여 정량적으로 산정하였고, 시간 종속형 공변수로 모형화된 공변수는 시간에 따른 공변수 영향의 변화를 분석하였다. 순차적 파손사건에 대한 비례위험모형의 구축 결과 생존시간군(STG) I의 기저위험률은 매설 후 대략 450개월까지는 파손 위험률이 '0'에 가까우나 그 이후로 급격히 증가하다가 매설 후 약 700개월에 이르러서는 약간 감소하고 약 850개월 이후에는 다시 급격히 증가한다. STG II의 기저위험률은 첫 번째 파손 후 약 300개월이 되면 위험률이 급격히 증가하는 것으로 나타났다. STG III $\sim$ STG VII의 기저위험률은 이차함수의 형태를 띄며, 특히 STG V, STG VI 및 STG VII의 기저위험률은 욕조형 곡선(bathtub curve)의 형태를 가진다. 각 생존시간군의 기저생존함수의 생존확률 '0.5'에 해당하는 기저중간생존시간에 대한 분석으로부터 파손횟수가 많아질수록 순차적 파손사건 사이의 경과시간은 감소하는 것으로 나타났다. 이러한 기저생존시간에 대한 경향은 관로의 파손횟수가 많아질수록 관로의 일반적은 내구성은 감소하기 때문인 것으로 분석된다.

• The Korean Journal of Applied Statistics
• /
• v.11 no.1
• /
• pp.163-175
• /
• 1998

The change of hazard rates at some unknown time point has been the interest of many statisticians. But it was restricted to the constant hazard rates which correspond to the exponential distribution. In this paper we generalize the change-point model in which any specific functional forms of hazard rates are net assumed. The assumed model includes various types of changes before and after the unknown time point. The Nelson estimator of cumulative hazard function is introduced. We estimate the change-point maximizing slope changes of Nelson estimator. Consistency and asymptotic distribution of bootstrap estimator are obtained using the martingale theory. Through a Monte Carlo study we check the performance of the proposed method. We also explain the proposed method using the Stanford Heart Transplant Data set.

• Journal of Korea Water Resources Association
• /
• v.42 no.6
• /
• pp.445-455
• /
• 2009

In this paper proportional hazards models for the first through seventh break of 150 mm cast iron pipes in a case study area are established. During the modeling process the assumption of the proportional hazards for covariates on the hazards is examined to include the time-dependent covariate terms in the models. As a result, the pipe material/joint type and the number of customers are modeled as time-dependent for the first failure, and for the second failure only the number of customers is modeled as time-dependent. From the analysis on the baseline hazard functions the failure hazards are found to be generally increasing for the first and second failure, while the hazards of the third break and beyond showed a form of a bath-tub. Furthermore, the changes in the baseline hazard rates according to the time and number of break reflect that the general condition of the pipes is deteriorating. The factors causing pipe break and their effects are analyzed based on the estimated regression coefficients and their hazard ratios, and the constructed models are verified using the deviance residuals of the models.

• The Journal of the Convergence on Culture Technology
• /
• v.6 no.1
• /
• pp.455-461
• /
• 2020

Electronic payment system using Internet banking is a very important application for users of e-commerce environment. With rapidly growing use of fintech applications, the risk and damage caused by malicious hacking or identity theft are getting significant. To prevent the damage, fraud detection system (FDS) calculates the risk of the electronic payment transactions using user profiles including types of goods, device status, user location, and so on. In this paper, we propose a new risk calculation method using relative location of users such as SSID of wireless LAN AP and MAC address. Those relative location information are more difficult to imitate or copy compared with conventional physical location information like nation, GPS coordinates, or IP address. The new method using relative location and cumulative user characteristics will enable stronger risk calculation function to FDS and thus give enhanced security to electronic payment systems.

• The Korean Journal of Applied Statistics
• /
• v.19 no.1
• /
• pp.163-170
• /
• 2006

The argument concerning the choice of effect measure for epidemiologic data or clinic data has been renewed. But the relationships among effect measures can be confusing if effect measures are expressed by conventional mathematical functions alone. In this article, risk difference(RD), relative risk(RR), and odds ratios(OR) for binary data are presented by radar diagram instead of mathematical functions and the relationships among them are showed using radar diagram. This radar diagram is offered flexible conceptual tool to understand effect measures, DR, RR, and OR for binary data.

• Journal of Bio-Environment Control
• /
• v.25 no.4
• /
• pp.308-319
• /
• 2016

In order to set the outdoor weather conditions to be applied to the design standard of the greenhouse heating and cooling system, outdoor air temperature and heating degree-hour for heating design, dry bulb temperature, wet bulb temperature and solar irradiance for cooling design were analyzed and presented. For every region in Korea, we used thirty years from 1981 to 2010 hourly weather data for analysis, which is the current standard of climatological normal provided by KMA. Since the use of standard weather data is limited, design weather conditions were obtained using the entire weather data for 30 years, and the average value of the entire data period was presented as a design standard. The design weather data with exceedance probability of 1, 2.5, and 5% were analyzed by the TAC method, and we presented the distribution map with exceedance probability of 1% for heating and 2.5% for cooling which are recommended by design standards. The changes of maximum heating load, seasonal heating load and maximum cooling load were examined by regions, exceedance probabilities, and setpoint temperatures. The proposed outdoor design conditions can be used not only directly for the greenhouse heating and cooling design, but also for the reinforcement of heating and cooling facilities and the establishment of energy saving measures. Recently, due to the climate change, sweltering heat in summer and abnormal temperature in winter are occurring frequently, so we need to analyze weather data periodically and revise the design standard at least every 10 years cycle.

• The Korean Journal of Applied Statistics
• /
• v.37 no.2
• /
• pp.225-237
• /
• 2024

In the analysis of competing risks data, some of covariates may not be fully observed for some subjects. In such cases, excluding subjects with missing covariate values from the analysis may result in biased estimates and loss of efficiency. In this paper, we studied multiple imputation and the augmented inverse probability weighting method for regression parameter estimation in the cause-specific proportional hazards model with missing covariates. The performance of estimators obtained from multiple imputation and the augmented inverse probability weighting method is evaluated by simulation studies, which show that those methods perform well. Multiple imputation and the augmented inverse probability weighting method were applied to investigate significant risk factors for the risk of death from breast cancer and from other causes for breast cancer data with missing values for tumor size obtained from the Prostate, Lung, Colorectal, and Ovarian Cancer Screen Trial Study. Under the cause-specific proportional hazards model, the methods show that race, marital status, stage, grade, and tumor size are significant risk factors for breast cancer mortality, and stage has the greatest effect on increasing the risk of breast cancer death. Age at diagnosis and tumor size have significant effects on increasing the risk of other-cause death.