• Journal of the Korean Data and Information Science Society
• /
• v.14 no.3
• /
• pp.673-685
• /
• 2003

In this paper we consider the problem of identifying and testing outliers in linear regression. First we consider the problem for testing the null hypothesis of no outliers. A test based on the ratio of two residual scale estimates is proposed. We show the asymptotic distribution of the test statistics by Monte Carlo simulation and investigate its properties. Next we consider the problem of identifying the outliers. A forward sequential procedure using the suggested test is proposed and shown to perform fairly well. Unlike other forward procedures, the present one is unaffected by masking and swamping effects because the test statistic is based on robust scale estimate.

• The Korean Journal of Applied Statistics
• /
• v.22 no.5
• /
• pp.1033-1046
• /
• 2009

In new drug development studies or clinical trials, zero-dose control is needed in general to determine the lowest dose level for a new drug which can act with our bodies. When the lowest dose level compared with zero-dose control has significant difference in effect, it is referred as minimum effective dose(MED). We propose, in this paper, parametric sequential test using updated control to identify the minimum effective dose(MED) level. Monte Carlo Simulation is adapted to examine the power and experimental significance levels of the proposed method with other methods.

• Journal of Aerospace System Engineering
• /
• v.11 no.4
• /
• pp.22-29
• /
• 2017

A study is performed for the effective detection method of a fault which is occurred during operation in a small turbojet engine with non-linear characteristics used by unmanned air vehicle. For this study the non-linear dynamic model of the engine is derived from transient thermodynamic cycle analysis. Also for inducing real operation conditions the controller is developed associated with unscented Kalman filter to estimate noises. Sequential probability ratio test is introduced as a real time method to detect a fault which is manipulated for simulation as a malfunction of rotational speed sensor contaminated by large amount of noise. The method applied to the fault detection during operation verifies its effectiveness and high feasibility by showing good and definite decision performances of the fault.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.1
• /
• pp.1-5
• /
• 2023

Most of the production process is performed in a cleanroom in the case of facilities that produce semiconductor chips or display panels. Therefore, environmental management of cleanrooms is very important for product yield and quality control. Among them, airborne particles are a representative management item enough to be the standard for the actual cleanroom rating, and it is a part of the Fab or Facility monitoring system, and the sequential particle monitoring system is mainly used. However, this method has a problem in that measurement efficiency decreases as the length of the sampling tube increases. In addition, a statistically significant test of deterioration in efficiency has rarely been performed. Therefore, in this study, the statistically significant test between the number of particles measured by InSitu and the number of particles measured for each sampling tube ends(Remote). Through this, the efficiency degradation problem of the sequential particle monitoring system was confirmed by a statistical method.

• Animal cells and systems
• /
• v.5 no.4
• /
• pp.313-317
• /
• 2001

The sequential assessment model describes a fight between two conspecific as an ongoing statistical sampling process, which makes it possible to predict fight length or repetition number of a behavioral element depending on relative RHP (resource holding potential: e.g. weight or fighting ability). We staged contests between males of common freshwater gobies to test some predictions of this model. Fights proceeded in a consistent sequence of phases. Most contests began with two contestants adopting lateral display, and then escalated to intense physical contacts. The length of contests was negatively correlated with weight difference between the contestants. The duration of complete phases was, however, independent of weight, and the prior information gained during complete phases did not appear to affect subsequent phases of the fight. Our results show that the contests of common freshwater gobies are well predicted by the sequential assessment model.

• Journal of the Korean Geotechnical Society
• /
• v.19 no.5
• /
• pp.7-14
• /
• 2003

In the back analysis of braced excavation based on displacement measurements, the accumulated errors of abnormal displacement caused by unexpected loads are usually observed. To minimize such errors, in this study, displacement increment was used fur back analysis rather than displacement for the conventional method. That is, the increment of sequential displacement is used as the objective function and the back analysis program is developed by taking advantage of sequential linear programming method by modified method of feasible directions. Based on small scale model test, the analysis of proposed method is verified.

• Journal of Mechanical Science and Technology
• /
• v.16 no.5
• /
• pp.710-719
• /
• 2002

This study investigates a nonlinear inverse convection problem for a laminar-forced convective flow between two parallel plates. The upper plate is exposed to unknown heat flux while the lower plate is insulated. The unknown heat flux is determined using temperature measured on the lower plate. The thermophysical properties of the fluid are temperature dependent, which renders the problem nonlinear. The sequential gradient method is applied to this nonlinear inverse problem in order to solve the problem efficiently. The function specification method is incorporated to stabilize the sequential estimation. The corresponding adjoint formalism is provided. Accuracy and stability have been examined for the proposed method with test cases. The tendency of deterministic error is investigated for several parameters. Stable solutions are achieved eve]1 with severely impaired measurement data.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.22 no.4
• /
• pp.153-158
• /
• 2008

When we design the control system used Programmable Logic Controller(PLC), we program the main algorithm by Ladder Diagram(LD) in many standard language. But if we use an Sequential Function Chart(SFC), we is easy to know the sequential flow of control, to maintenance the controller and to describe a program. Even though a SFC have many merits, A SFC have some problems like a condition and a function of self test. Therefore this thesis show an efficient function of self test and we confirmed its feasibility through actual example.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.53 no.1
• /
• pp.22-27
• /
• 2004

BiCMOS circuit consist of CMOS part which constructs logic function, and bipolar part which drives output load. In BiCMOS circuits, transistor stuck-open faults exhibit delay faults in addition to sequential behavior. In this paper, proposes a method for efficiently generating test pattern which detect stuck-open in BiCMOS circuits. In proposed method, BiCMOS circuit is divided into pull-up part and pull-down part, using structural property of BiCMOS circuit, and we generate test pattern using set theory for efficiently detecting faults which occured each divided blocks.

• Journal of the Korea Society of Computer and Information
• /
• v.4 no.1
• /
• pp.28-37
• /
• 1999

In this paper. we developed a delay test architecture and test procedure for clocked 5-bit asynchronous counter circuit based on boundary scan architecture. To develope, we analyze the problems of conventional method on delay test for clocked sequential circuit in boundary scan architecture. This paper discusses several problems of delay test on boundary scan architecture for clocked sequential circuit. Conventional test method has some problems of improper capture timing, of same pattern insertion, of increase of test time. We suggest a delay test architecture and test procedure, is based on a clock count-generation technique to generate continuous clocks for clocked input of CUT. The simulation results or 5-bit counter shows the accurate operation and effectiveness of the proposed delay test architecture and procedure.