• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11b
• /
• pp.382-384
• /
• 2005

소프트웨어 테스팅에서 테스트 스위트(suite)의 수를 줄이면서도 테스트 커버리지나 오류검출에 있어서 효과적인 방법을 찾기 위한 많은 연구가 시도되어 왔다. 안정성이 높은 설계가 되도록 설계 조건을 결정하기 위하여 제어인자들의 직교 배열을 사용하는 시험방법인 교호작용 테스팅 기법이 소프트웨어의 테스트에서도 효과가 매우 높다는 것이 실험으로 증명되고 있다. 소프트웨어는 상대적으로 몇 안 되는 조건들의 조합들로 오류가 발생할 가능성이 높다는 특징을 가지고 있다. 따라서, 파라미터 간 교호작용 강도 t를 갖는 t-way 테스팅을 통해 효과적으로 테스트 스위트를 줄이면서 많은 오류를 검출할 수 있다. 그러나 t값을 증가시키면 테스트 스위트의 수가 늘어난다는 단점이 발생한다. 또한, 어떠한 파라미터들이 서로 교호작용을 일으키는 것인지 알수 없는 상황을 가정한 단순한 교호작용 테스팅은 자칫 서로 관련 없는 파라미터들도 교호작용 테스팅에 참여하기 때문에 테스트의 낭비가 있을 수 있다는 단점이 있다. 이에, 본 논문에서는 소프트웨어의 입력과 출력간의 관계를 바탕으로 시스템 I-O관계도를 작성한 후, 이를 바탕으로 각 출력에 대한 교호작용을 일으킬 수 있는 파라미터를 중심으로 테스트 스위트를 생성하는 Selective Covering Array를 제안한다.

• The Korean Journal of Applied Statistics
• /
• v.7 no.2
• /
• pp.201-211
• /
• 1994

Ignoring interaction effects has been pointed out to be one of serious drawbacks in analysis of the parameter designs which are constructed by using orthogonal arrays. In this paper a detecting procedure of influential 2-factor inteactions with minimum expeimental runs is described, when each contrl factor has two levels. The presented method is based on the near orthogonal arrays which are very similar to orthogonal arrays in the statistical structure. And those arrarys are the same as trace-optimal balanced saturated two-level fractional factorial designs of resolution V.

• Journal of Korea Water Resources Association
• /
• v.51 no.spc
• /
• pp.1067-1078
• /
• 2018

Water resources projection typically consists of several stages including emission scenarios, global circulation models (GCMs), downscaling techniques, and hydrological models, and each stage is a source of total uncertainty in water resources projection. Several studies proposed methods to quantify the relative contribution of each stage to total uncertainty, and we call such analysis uncertainty decomposition. Uncertainty decomposition enables us to investigate the stages yielding large uncertainties and to establish the uncertainty reduction plan that reflects them. Interactions between stages is one of the important issues to be considered in uncertainty decomposition. This study suggests a new uncertainty decomposition method considering interaction effect. The proposed method has an advantage of decomposing the total uncertainty to the uncertainty from each stage considering both the main and interactions effects. We apply the proposed method to streamflow projection for Chungju Dam basin. The results show that the uncertainties from the main effects are larger than the uncertainties from interaction effects in both summer and winter. Using the proposed uncertainty decomposition method, we show that the GCM stage is the largest source of the total uncertainty in summer and the downscaling technique stage is the one in winter among the following four stages: emission scenarios, GCMs, downscaling techniques, and hydrological models.

• Journal of Korean Society for Quality Management
• /
• v.45 no.1
• /
• pp.55-64
• /
• 2017

Purpose: A large number of attributes with mixed levels are often considered in the conjoint analysis. The respondents may have difficulty with scoring their preferences accurately because of many attribute items involved in each survey question. We research on the technique for reducing the number of attribute items. Methods: In order to reduce the number of attribute items in a survey question, we make a new attribute by merging two original attributes. A 'No question' option is also included as a new level in a merged attribute. Results: We propose BIB $6^4$ design in the case where we have four attributes with 2 levels and 3 levels, respectively and then analyze all the respondents survey data generated by the repeated simulation study in order to compare various model selection methods. Conclusion: How to reduce the number of attribute items is proposed and how to design and analyze the survey data are illustrated.

• The Korean Journal of Applied Statistics
• /
• v.28 no.5
• /
• pp.951-963
• /
• 2015

The conjoint analyst in marketing are anxious to know whether there exist synergy or antagonistic effects between two attributes. That is to say, they are interested in estimating the main effects as well as the two factor interaction effects.We research the design of survey questionnaire so that all the main effects and two factor interaction effects are estimable by employing the resolution V balanced Incomplete Block Fractional Factorial Design. We screen vital few effects, find the proper model and obtain information for efficient concepts optimization by analyzing all respondents survey data.

• KOREAN POULTRY JOURNAL
• /
• v.14 s.153
• /
• pp.102-105
• /
• 1982

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.37 no.4
• /
• pp.320-328
• /
• 1992

The soybean yield forcasting models based on climatic elements in six locations were estimated by the STEPWISE/MAXR, Cp statistics and GLM procedure of SAS. The climatic elements were aerial temperature, sunshine hours and precipitation from May to October in 20 years. The investigated six locations were Chunchon, Suwon, Cheongju, Kwangju, Iri and Jinju. The important climatic elements for main effects in Chunchon model were August sunshine hours-linear term, August precipitation-quadratic. June temperature to August precipitation and May temperature to August precipitation were interaction terms. The quadratic August precipitation was assumed to be related to yield in Chunchon. The main effects of Suwon were linear-June temperature, quadratic June sunshine hours and June precipitation. These terms affected yields negatively. The main effects of Cheongju were linear June temperature and quadratic August precipitation. May temperature to June precipitation, July to August precipitations were interactions. The main effects of Kwangju were linear July precipitation, quadratic June temperature and July precipitation. June to July sunshine hours of interaction terms influenced yield negatively. The main effects of Iri were linear May sunshine hours, quadratic May and July sunshine hours. May temperature to May precipitation and June to July precipitations affected yields negatively. The main effects of Jinju were linear June and August precipitations. August temperature to August sunshine hours, June sunshine hours to July precipitation and June to August precipitation were interactions. In linear terms, June and August precipitations and, in interactions, August to August sunshine hours were negative efficacies respectively. The included year variables in Chunchon, Suwon, Kwangju, and Jinju model building were recognized as a linear trend based on an assumption that the technological factors have improved through times.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.43 no.4
• /
• pp.592-603
• /
• 2014

We developed various recipes of turmeric powder (Curcuma longa L.) added to milk bread and assessed the individual effects of seven ingredients [milk ($X_1$), turmeric powder ($X_2$), bread improver ($X_3$), fresh yeast ($X_4$), butter ($X_5$), sugar ($X_6$), and salt ($X_7$)] as well as the 2-way interaction effects of the ingredients on the sensory characteristics of breads using fractional factorial design method. The center and end points of each component were determined via literature review and multiple test baking. Seven trained sensory test panels evaluated the outside appearance (OA), inside appearance (IA), and flavor & texture (FT) of 38 breads using 46 items of sensory evaluation. Findings are as follows: for the OA, $X_1$ (P<0.05) and $X_4$ (P<0.0001) exhibited significant individual effects, whereas $X_1*X_7$, $X_2*X_5$, $X_3*X_6$, and $X_4*X_6$ indicated significant interaction effects (P<0.05). For the IA, $X_1$ (P<0.0001), $X_4$ (P<0.0001), $X_6$ (P<0.05), $X_2*X_4$ (P<0.05), and $X_3*X_6$ (P<0.01) showed individual and interaction effects, respectively. For the FT, $X_1$ and $X_2$ showed the most significant individual effect (P<0.0001), followed by $X_4$, $X_5$ and $X_6$ (P<0.05) in descending order. $X_4*X_7$ indicated the only significant interaction effect. We computed the magnitudes of the 2-way interaction effects of the ingredients with a distinct emphasis. Model equations predicting the levels of the ingredient effects on the breads were also provided via regression analyses. In summation, $X_4$ appeared to be the most significant component affecting the sensory characteristics based on its individual and 2-way interaction effects. Further, $X_6$, $X_1$, $X_2$, and $X_5$ indicated both individual and interaction effects. $X_3$ and X7 showed only interaction effects. The center point effect appeared to be unequivocal for whole sensory characteristics. Findings of the present study may provide insights into the selection of ingredients to derive an optimal model for turmeric powder-added bread using the response surface method hereafter.

• Korean journal of applied entomology
• /
• v.14 no.4 s.25
• /
• pp.205-208
• /
• 1975

Effect of light and temperature on the sporulation and mycelial growth by 4 strains of Mycosphaerella fragariae was examined. The light source was day light fluorescent lamps with an intensity of 2000 Lux at the colony surface. The treatment were constant light, alternating light(10hr) and darness (14h.) and constant darkness af $15^{\circ}C,\; 22^{\circ}C\; and\; 30^{\circ}C$ 1. Of the 4 strains tested, alternating light and darkness at $22^{\circ}C\;and\;15^{\circ}C$ gave the greatest sroducedtion, constant light in intermediate, constant darkness in minimum, but darkness at $30^{\circ}C$ produced abundant conidia over that obtained in the other two treatment. Differences in number of sporulation due to light, strains and the interaction of these two were all statistically significant. 2. There were not on significant differences in amount of mycelial growth due to light, but amount of mycelial growth of$S_3$ strain were the nest abundant among the 4 strains.

• Journal of The Korean Association For Science Education
• /
• v.37 no.4
• /
• pp.659-668
• /
• 2017

This study was conducted to identify whether there is a significant difference in the developmental stages of the evolution concept depending on the major and the country. For this study, college students (both biology major and non-biology major) in Korea and the United States answered evolution concept assessment developed by Ha (2007). The data were analyzed based on the method of developed by Ha (2016). As a result of this study, Cronbach alpha for internal consistency reliability and MNSQ indices for item response fitness of the ordered multiple choice method providing 0 to 4 point scores to creationism, teleology, intentionality, use/disuse, and natural selection respectively, satisfied the benchmark. The level of Korean college students' evolution concepts was located in use/disuse and the level of USA college students' evolution concepts was located in teleology. No interaction effect was found between the difference of conceptual progressions of evolution concept by major and country and evolution contexts (human, animal, and plant). There was a significant relationship between developmental stages of evolution concept and their religiosity. Based on the results of this study, we discuss developmental characteristics of evolutionary concepts.