• 소프트웨어공학소사이어티 논문지
• /
• 제28권1호
• /
• pp.23-28
• /
• 2019

최근 이미지 처리 응용을 위한 심층 신경망 모델의 효과적 테스팅을 위해 해당 모델이 올바르게 예측하지 못하는 코너 케이스에 해당하는 행동을 보이는 데이터를 자동 생성하는 연구가 활발히 진행되고 있다. 본 논문은 문장 분류 심층 신경망 모델에 기반하고 있는 버그 담당자 자동 배정 시스템의 테스트를 위해 입력 데이터인 버그 리포트의 내용에서 임의의 단어를 선택해 동의어로 변형하는 테스트 데이터 생성기법을 제안한다. 그리고 제안하는 테스트 데이터 생성 기법을 사용한 경우와 기존의 차이 유발 테스트 데이터 생성 기법을 사용했을 경우를 다양한 뉴런 기반 커버리지를 중심으로 비교 평가한다.

• Journal of the Korean Statistical Society
• /
• 제30권4호
• /
• pp.645-660
• /
• 2001

In this paper we propose a goodness-of-fit test statistic for testing the (null) parametric model versus the (alternative) nonparametric model. Most of existing nonparametric test statistics are based on the residuals which are obtained by regressing the data to a parametric model. Our test is based on the bootstrap estimator of the probability that the smoothing parameter estimator is infinite when fitting residuals to cubic smoothing spline. Power performance of this test is investigated and is compared with many other tests. Illustrative examples based on real data sets are given.

• Journal of the Korean Data and Information Science Society
• /
• 제18권3호
• /
• pp.809-815
• /
• 2007

Since Bollerslev(1986), the GARCH model has been popular in analysing the volatility of the financial time series. In real data analysis, practitioners conventionally put the normal assumption on the innovation random variables of the GARCH model, which is often violated. In this paper, we analyse the domestic financial data based on the GARCH(1,1) model and among existing normality tests, perform the Jarque-Bera test based on the residuals. It is shown that the innovation based on the GARCH(1,1) model dose not follow the normality assumption.

• Journal of the Korean Data and Information Science Society
• /
• 제19권3호
• /
• pp.893-901
• /
• 2008

The disconfirmation paradigm is the earliest researched and the most deeply researched of all the paradigms in marketing. Disconfirmation paradigm deals with the influence of expectation, perceived product performance, and the discord between the two on consumer satisfaction. The GAP-Model is based on the disconfirmation paradigm that tries to understand the effect of the gap between before purchase expectations and after purchase perceptions of the product performance on dependent variables such as customer satisfaction. The purpose of this research is to test whether regression coefficients of a P-Model(performance only model), an E-Model(expectation only model) and GAP(P-E)-Model are equivalent in explaining service value and loyalty. The Chow's F-Test is used to test the excellence of the 3 models. As a result of comparison and analysis, P-Model showed more excellence of service value and loyalty than E-Model or GAP-Model.

• Elastomers and Composites
• /
• 제51권4호
• /
• pp.308-316
• /
• 2016

The rubber materials are widely used in automobile industry due to their capability of a large amount of elastic deformation under a force. Current trend of design process requires prediction of functional properties of parts at early stage. The behavior of rubber material can be modeled using strain energy density function. In this study, five different strain energy density functions - Neo-Hookean model, Reduced Polynomial $2^{nd}$ model, Ogden $3^{rd}$ model, Arruda Boyce model and Van der Waals model - were used to estimate the behavior of carbon black added natural rubber under uniaxial load. Two kinds of tests - uniaxial tension test and biaxial tension test - were performed and used to correlate the coefficients of the strain energy density function. Numerical simulations were carried out using finite element analysis and compared with experimental results. Simulation revealed that Ogden $3^{rd}$ model predicted the behavior of carbon added natural rubber under uniaxial load regardless of experimental data selection for coefficient correlation. However, Reduced Polynomial $2^{nd}$, Ogden $3^{rd}$, and Van der Waals with uniaxial tension test and biaxial tension test data selected for coefficient correlation showed close estimation of behavior of biaxial tension test. Reduced Polynomial $2^{nd}$ model predicted the behavior of biaxial tension test most closely.

• Journal of the Korean Data and Information Science Society
• /
• 제15권3호
• /
• pp.645-654
• /
• 2004

In this paper, we develop the Bayesian test procedure for the intraclass correlation coefficient in the unbalanced one-way random effect model based on the reference priors. That is, the objective is to compare two nested model such as the independent and intraclass models using the factional Bayes factor. Thus the model comparison problem in this case amounts to testing the hypotheses $H_1:\rho=0$ versus $H_2:{\rho}{\neq}0$. Some real data examples are provided.

• 한국구조물진단유지관리공학회 논문집
• /
• 제23권4호
• /
• pp.53-60
• /
• 2019

이 연구에서는 정적 변위를 사용하는 유전자 알고리즘을 이용한 교량의 유한요소해석모델 개선 기법을 제안하며, PSC 거더교를 대상으로 한 실증시험 데이터를 이용하여 제안된 방법을 검증하였다. 실증 재하시험으로 정적재하시험과 의사정적재하시험을 수행하였으며, 각 재하시험의 계측 데이터를 이용하여 대상교량의 유한요소해석모델 개선을 진행하였다. 최종적으로 의사정적재하시험의 계측 데이터를 통해 개선된 모델을 이용하여 공용 내하력평가를 수행하였다. 내하력평가에는 현 도로교설계기준과 구 도로교설계기준, AASHTO LRFD의 설계 활하중을 이용하였으며, 각 설계기준 별 내하력평가 결과를 비교하였다.

• Journal of the Korean Data and Information Science Society
• /
• 제24권6호
• /
• pp.1465-1475
• /
• 2013

The chi-square type test statistic is the most commonly used test in terms of measuring testing goodness-of-fit for multinomial logistic regression model, which has its grouped data (binomial data) and ungrouped (binary) data classified by a covariate pattern. Chi-square type statistic is not a satisfactory gauge, however, because the ungrouped Pearson chi-square statistic does not adhere well to the chi-square statistic and the ungrouped Pearson chi-square statistic is also not a satisfactory form of measurement in itself. Currently, goodness-of-fit in the ordinal setting is often assessed using the Pearson chi-square statistic and deviance tests. These tests involve creating a contingency table in which rows consist of all possible cross-classifications of the model covariates, and columns consist of the levels of the ordinal response. I examined goodness-of-fit tests for a proportional odds logistic regression model-the most commonly used regression model for an ordinal response variable. Using a simulation study, I investigated the distribution and power properties of this test and compared these with those of three other goodness-of-fit tests. The new test had lower power than the existing tests; however, it was able to detect a greater number of the different types of lack of fit considered in this study. I illustrated the ability of the tests to detect lack of fit using a study of aftercare decisions for psychiatrically hospitalized adolescents.

• 한국신뢰성학회지:신뢰성응용연구
• /
• 제15권2호
• /
• pp.90-100
• /
• 2015

Reliability growth rate (or reliability growth curve slope) have the two cases of trend as a constant or changing one during the reliability growth testing. The changing case is very common situation. The reasons of reliability growth rate changing are that the failures to follow the NHPP (None-Homogeneous Poisson Process), and the solutions implemented during test to break out other problems or not to take out all of the root cause permanently. If the changing were big, the "Goodness of Fit (GOF)" of reliability growth curve to test data would be very low and then reduce the accuracy of assessing result with test data. In this research, we are using Duane model and AMSAA model for assessing test data and projecting the reliability level of complex and repairable system as like construction equipment and vehicle. In case of no changing in reliability growth rate, it is reasonable for reliability engineer to implement the original Duane model (1964) and Crow-AMSAA model (1975) for the assessment and projection activity. However, in case of reliability growth rate changing, it is necessary to find the method to increase the "GOF" of reliability growth curves to test data. To increase GOF of reliability growth curves, it is necessary to find the proper parameter calculation method of interesting reliability growth models that are applicable to the situation of reliability growth rate changing. Since the Duane and AMSAA models have a characteristic to get more strong influence from the initial test (or failure) data than the latest one, the both models have a limitation to contain the latest test data information that is more important and better to assess test data in view of accuracy, especially when the reliability growth rate changing. The main objective of this research is to find the parameter calculation method to reflect the latest test data in the case of reliability growth rate changing. According to my experience in vehicle and construction equipment developments over 18 years, over the 90% in the total development cases are with such changing during the developing test. The objective of this research was to develop the newly assessing method and the process for GOF level increasing in case of reliability growth rate changing that would contribute to achieve more accurate assessing and projecting result. We also developed the new evaluation method for GOF that are applicable to the both models as Duane and AMSAA, so it is possible to compare it between models and check the effectiveness of new parameter calculation methods in any interesting situation. These research results can reduce the decision error for development process and business control with the accurately assessing and projecting result.

• 한국군사과학기술학회지
• /
• 제23권4호
• /
• pp.389-398
• /
• 2020

Test cases in the existing weapon system software were created manually by the tester analyzing the test items defined in the software integration test procedure. However, existing test case generation method has two limitations. First, the quality of test cases can vary depending on the tester's ability to analyze the test items. Second, excessive time and cost may be incurred in writing test cases. This paper proposes a method to automatically generate test cases based on the requirements model and specifications to overcome the limitations of the existing weapon system software test case generation. Generate test sequences and test data based on the use case event model, a model representing the requirements of the weapon system software, and the use case specification specifying the requirements. The proposed method was applied to 8 target models constituting the avionics control system, producing 30 test sequences and 8 test data.