• Journal of the Korean Statistical Society
• /
• v.25 no.3
• /
• pp.433-444
• /
• 1996

In contrast to the multiplicative risk model, the additive risk model specifies that the hazard function with covariates is the sum of, rather than product of, the baseline hazard function and the regression function of covariates. We, in this paper, propose a method for checking the adequacy of the additive risk model based on partial-sum of matingale residuals. Under the assumed model, the asymptotic properties of the proposed test statistic and approximation method to find the critical values of the limiting distribution are studied. Several real examples are illustrated.

• Journal of the Korean Data and Information Science Society
• /
• v.11 no.2
• /
• pp.201-215
• /
• 2000

The classical method for regression analysis is the least squares method. However, if the data contain significant outliers, the least squares estimator can be broken down by outliers. To remedy this problem, the robust methods are important complement to the least squares method. Robust methods down weighs or completely ignore the outliers. This is not always best because the outliers can contain some very important information about the population. If they can be detected, the outliers can be further inspected and appropriate action can be taken based on the results. In this paper, I propose a sequential outlier test to identify outliers. It is based on the nonrobust estimate and the robust estimate of scatter of a robust regression residuals and is applied in forward procedure, removing the most extreme data at each step, until the test fails to detect outliers. Unlike other forward procedures, the present one is unaffected by swamping or masking effects because the statistics is based on the robust regression residuals. I show the asymptotic distribution of the test statistics and apply the test to several real data and simulated data for the test to be shown to perform fairly well.

• Proceedings of the KIEE Conference
• /
• 1997.07b
• /
• pp.485-487
• /
• 1997

Schemes for detecting and diagnosing faults are presented. Faults are detected when residuals change significantly and thresholds are exceed. Two stage artificial neural networks are applied to diagnose faults. The idealized steady state patterns of residuals are defined and learned by ANNs using back propagation algorithm. The first stage ANN is trained to classify the subsystem in which the various faults are located. The first stage ANN could be also used to detect faults with threshold, checking. The second stage ANNs are trained to discriminate the specific cause of a fault at the subsystem level.

• The Korean Journal of Applied Statistics
• /
• v.9 no.1
• /
• pp.75-89
• /
• 1996

This paper proposes a goodness-of-fit test for checking the adequacy of the additive risk model with a binary covariate. The test statistic is based on martingale residuals, which is the extended form of Wei(1984)'s test. The proposed test is shown to be consistent and asymptotically normally distributed under the regularity conditions. Furthermore, the test procedure is illustrated with two set of real data and the results are discussed.

• Journal of Korean Society of Water and Wastewater
• /
• v.25 no.5
• /
• pp.777-782
• /
• 2011

The residual treatment facilities in WTP(water treatment plant) play an important role in solid-liquid separation. At present, it is difficult to solve problems related with thickening and dewatering of WTP sludge, and discharging waste water to river. The quantity of residuals generated from water treatment plants depends upon the raw water quality, dosage of chemicals used, performance of the treatment process, method of sludge removal, efficiency of sedimentation, and backwashing frequency. Sludge production by the physical separation of SS(Suspended Solid) occurs under quiescent conditions in the primary clarifier, where SSs are allowed to settle and to consolidate on the clarifier bottom. Raw primary sludge results when the settled solids are hydraulically removed from the tank. In this study, Drawing characteristics of the sludge thickening in the thickener of Water Treatment Plants was simulated by Using CFD(Computational Fluid Dynamics.

• Journal of the Korean Society of Safety
• /
• v.29 no.4
• /
• pp.73-77
• /
• 2014

In the field of fault diagnosis, the deviations from normal operating conditions are monitored to identify the type of faults and find their root causes. One of the most representative methods is the statistical approaches, due to a large amount of advantages. However, ambiguous diagnosis results can be generated according to fault magnitudes, even if the same fault occurs. To tackle this issue, this work proposes principal component analysis (PCA) based method with qualitative information. The PCA model is constructed under normal operation data and the residuals from faulty conditions are calculated. The significant changes of these residuals are recorded to make the information for identifying the types of fault. This model can be employed easily and the tasks for building are smaller than these of other common approaches. The efficacy of the proposed model is illustrated in Tennessee Eastman process.

• Journal of Korean Society for Quality Management
• /
• v.26 no.1
• /
• pp.126-142
• /
• 1998

Traditional statistical process control(SPC) assumes that consective observations from a process are independent. In industrial practice, however, observations are ofter serially correlated. A common a, pp.oach to building control charts for autocorrelatd data is to a, pp.y classical SPC to the residuals from a time series model fitted. Unfortunately, one cannot completely escape the effects of autocorrelation by using charts based on residuals of time series model. For the detection of variance change in the process we propose a CUSUM of squares control chart which does not require the model identification. The proposed CUSUM of squares chart and the conventional control charts are compared by a Monte Carlo simulation. It is shown that the CUSUM of squares chart is more effective in the presence of dependency in the processes.

• Communications for Statistical Applications and Methods
• /
• v.30 no.4
• /
• pp.423-435
• /
• 2023

The Jarque and Bera (1980) statistic is one of the well known statistics to test univariate normality. It is based on the sample skewness and kurtosis which are the sample standardized third and fourth moments. Desgagné and de Micheaux (2018) proposed an alternative form of the Jarque-Bera statistic based on the sample second power skewness and kurtosis. In this paper, we generalize the statistic to a multivariate version by considering some data driven directions. They are directions given by the normalized standardized scaled residuals. The statistic is a modified multivariate version of Kim (2021), where the statistic is generalized using an empirical standardization of the scaled residuals of data. A simulation study reveals that the proposed statistic shows better power when the dimension of data is big.

• Journal of The Korean Astronomical Society
• /
• v.29 no.spc1
• /
• pp.461-462
• /
• 1996

• Journal of Korean Society of Water and Wastewater
• /
• v.20 no.5
• /
• pp.653-659
• /
• 2006