• Communications for Statistical Applications and Methods
• /
• v.31 no.6
• /
• pp.703-708
• /
• 2024

With the development of high-throughput technologies for producing genomic data, more advanced statistical methods such as regularization and machine learning techniques have been adapted to survival analysis. However, the clinical information such as age, gender and medical history plays a critical role in constructing a survival prediction model. Machine learning technique such as the support vector machine (SVM) can improve the predictability of the survival model by using the available clinical information. When implementing SVM for a predictive survival model, the clinical kernel was proposed by Daemen et al. by equalizing the influence of clinical variables and taking account of the range of these variables. However, this clinical kernel uses the same weight for all clinical variables without considering the different effect of those variables on the survival time. In this study, we proposed a simple kernel, called ensemble kernel, by combining a clinical kernel with model fitting. Since the proposed ensemble kernel is based on model fitting, two different kernels are considered by using either Cox model or accelerated failure time (AFT) model. We compare the performance of these two ensemble kernels with that of the linear kernel and the clinical kernel by the concordance index (C-index) using the four real data sets. While both linear and clinical kernels use all clinical variables in defining global kernels, the proposed two ensemble kernels can use only significant variables from either a Cox model or an AFT model. The comparative result shows that the proposed two ensemble kernels perform similarly as the existing clinical kernel does and the performance of four kernels vary according to data sets.

• Geomechanics and Engineering
• /
• v.38 no.5
• /
• pp.529-539
• /
• 2024

A high-level nuclear waste (HLW) repository is designed for the long-term disposal of high-level waste. Positioned at depths of 500-1000 meters, it offers an alternative to the insufficient storage space for spent fuels, providing a long-term solution. High-level waste emits heat and radiation, causing structural deterioration, including strength reduction and cracks. Therefore, the use of piezoelectric sensors for structural health monitoring is essential for evaluating the safety of the structure over time. Unlike other structures, the HLW repository restricts human access after the disposal of HLW, rendering sensor replacement impossible. Therefore, it is necessary to assess both the lifespan and suitability of sensors under the disposal conditions in the HLW repository. This study employed an accelerated life test (ALT) to assess the sensor's lifespan under disposal conditions. Failure modes, failure mechanisms, and operational limits were analyzed through accelerated stress test (AST). Additionally, the parameters of the Weibull life probability distribution and the Arrhenius accelerated life model were estimated through statistical methods, including the likelihood ratio test, maximum likelihood estimation, and hypothesis testing. Results confirmed that the sensor's lifespan decreases significantly with the increase in the temperature limit of the HLW repository. The findings of this study can be used for improving sensor lifespan through shielding, development of alternative sensors, or lifespan evaluation of alternative monitoring sensors.

• International Journal of Reliability and Applications
• /
• v.15 no.2
• /
• pp.85-98
• /
• 2014

Accelerated life testing (ALT) is a well famous technique in life testing and reliability studies, this is particularly used to induce so high stress leading to failure of the highly reliable units quickly under stipulated duration of time. The step-stress ALT is one of the systematic experimental strategy of ALT applied to fail the units in steps. In this article we focus on two important issues (i) necessity of life tests at higher steps with relevant causes (ii) to develop a new optimum test plan for 3-step SSALT under the modified cumulative exposure model proposed by Khamis and Higgins (1998). It is assumed that the lifetime of test units follows Rayleigh distribution and its scale parameter at constant stress level is assumed to be a log-linear function of the stress. The maximum likelihood estimates of the parameters involved in the step-stress ALT model are obtained. A simulation study is performed for numerical investigation of the proposed new optimum plan 3-step, step-stress ALT. The necessity of the life test units at 3-step step-stress is also numerically examined in comparison to simple step-stress setup.

• Journal of Applied Reliability
• /
• v.4 no.2
• /
• pp.73-90
• /
• 2004

We performed the heat aging test to predict the useful lifetime of Elastomeric Bearing Chloroprene Rubber (CR) used for supporting bridge. During the test, we measured elongation that are influenced by temperature and aging time. The failure of a test piece is defined as the point at which the elongation reaches to 75% of the initial value. This failure criterion is based on KS F 4420: 1998 (Elastomeric Bearing for bridge). Through the accelerated heat aging test, we found that the Arrhenius relationship and the Weibull lifetime distribution are appropriate as the life-temperature relationship and lifetime distribution of the CR, respectively. Using the Arrhenius -Weibull model, the parameters of the model are estimated and the lifetime of the CR at use condition is predicted.

• Journal of the Korean Data and Information Science Society
• /
• v.8 no.2
• /
• pp.183-194
• /
• 1997

In accelerated life tests, the failure time of an item is observed under a high stress level and based on the time, the failure rates of items we estimated at the normal stress level. In this paper, when the mean of the prior distribution of a parameter is known in Weibull lifetime model with censored failure time data, we study various estimating methods to obtain the empirical Bayes estimator of a parameter from the empirical Bayes approach under the normal stress level by considering the fact that the Bayes estimator is the function of prior parameters and of the acceleration parameter representing the effect of acceleration. And we compare the performance of several empirical Bayes estimators of a parameter in terms of the Bayes risk.

• International Journal of Reliability and Applications
• /
• v.15 no.1
• /
• pp.23-50
• /
• 2014

This paper deals with formulation of optimum multi-objective modified step-stress accelerated life test (ALT) plan for Burr type-XII distribution under type-I censoring. Since it is impractical to estimate only one objective parameter after conducting costly ALT tests; also, it is not desirable to assume instantaneous changes in stress levels because of limited capacity of test equipments and the presence of undesirable failure modes, therefore, an optimum multi-objective modified step-stress ALT plan has been designed. The optimal test plan consists in determining the optimum low stress level and optimal time at which stress starts linearly increasing from low stress by minimizing the weighted sum of the asymptotic variances of the maximum likelihood estimator of quantile lifetimes at design constant stress. The method developed has been illustrated using an example. Sensitivity analysis has been carried out. Comparative study has also been done to highlight the merits of the proposed model.

• Journal of Applied Reliability
• /
• v.9 no.4
• /
• pp.275-290
• /
• 2009

This paper considers accelerated thermal degradation tests which are performed for rubber seal materials used for undersea tunnels constructed by immersion method. Three types of rubber seals are tested; rubber expansion seal, omega seal, and shock absorber hose. Main ingredient of rubber expansion seal is EPDM(Ethylene Propylene Diene Monomer) and that of both omega seal and shock absorber hose is SBR(Styrene Butadiene Rubber). The accelerated stress is temperature and an Arrhenius model is introduced to describe the relationship between the lifetime and the stress. From the accelerated degradation tests, dominant failure mode of the rubber seals is found to be the loss of elongation. The lifetime distribution and the service life of the rubber seals at use condition are estimated from the test results. The acceleration factor for three types of rubber seals are also investigated.

• International Journal of Reliability and Applications
• /
• v.3 no.4
• /
• pp.185-198
• /
• 2002

In the traditional life testing model, it is assumed that a certain number of identical items are tested under identical condition. This is due to statistical rather than practical considerations. The proportional hazards model can be used to develop a realistic approach to determine the performance of an item. That is also capable of modeling the failure rates of accelerated life testing when the covariates are applied stresses. The proportional hazards model is typically applied for a group of items to assess the importance of factors that may influence the reliability of an item. In this paper we considered the interarrival times of an item rather than the time to first failure for grouped items and provided the availability estimation for the determination of maintenance policy and overhaul time. In order to demonstrate the proposed approach, an example is presented.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.41 no.10
• /
• pp.997-1004
• /
• 2017

The life of elevator wire ropes is one of the most important characteristics of an elevator, which is closely related to the safety of users and its maintenance policy. It is not cost effective to measure the lifetime of elevator wire ropes during their use. In this study, the life estimation of elevator wire ropes (8x19W-IWRC) is considered using accelerated degradation test data. A bending fatigue tester is used to perform the accelerated degradation tests, incorporating the acceleration factor of tensile force. Assuming that the life of wire ropes is log-normally distributed, two life estimation methods are suggested and their results are compared. The first method estimates the life of wire ropes utilizing the accelerated life model with pseudo lives obtained from a linear regression model. The second method estimates the life using a logistic model based on failure probability.

• The Korean Journal of Applied Statistics
• /
• v.34 no.3
• /
• pp.411-425
• /
• 2021

Accelerated failure time (AFT) model represents a linear relationship between the log-survival time and covariates. We are interested in the inference of covariate's effect affecting the variation of survival times in the AFT model. Thus, we need to model the variance as well as the mean of survival times. We call the resulting model mean and variance AFT (MV-AFT) model. In this paper, we propose a variable selection procedure of regression parameters of mean and variance in MV-AFT model using penalized likelihood function. For the variable selection, we study four penalty functions, i.e. least absolute shrinkage and selection operator (LASSO), adaptive lasso (ALASSO), smoothly clipped absolute deviation (SCAD) and hierarchical likelihood (HL). With this procedure we can select important covariates and estimate the regression parameters at the same time. The performance of the proposed method is evaluated using simulation studies. The proposed method is illustrated with a clinical example dataset.