• 한국지능시스템학회논문지
• /
• 제23권4호
• /
• pp.366-370
• /
• 2013

설명변수와 반응변수 사이의 통계적 관계를 설명하기 위해 사용되는 회귀모형을 분석하는 방법을 회귀분석이라 한다. 본 논문에서는 독립변수와 종속변수에 대한 퍼지관계를 표현하는 퍼지회귀모형를 추정하기 위하여 이상치에 민감하지 않은 로버스트한 추정량인 Theil방법을 소개한다. Theil방법은 설명변수와 반응변수의 ${\alpha}$-수준집합의 각 성분으로 구성된 집합에서 선택한 임의의 두 쌍 자료로부터 계산된 변화율의 중위수를 두 변수에 대한 변화량의 추정량으로 간주한다. 본 논문에서 제안된 Theil방법이 최소자승법을 이용하여 추정된 퍼지회귀모형보다 더 정확할 수 있음을 예제를 통하여 확인한다.

• 한국콘텐츠학회논문지
• /
• 제18권3호
• /
• pp.557-567
• /
• 2018

본 연구는 코스닥 IPO시장에서 공모주식 청약 시 일반투자자의 배정비율을 20%이상으로 규정한 2004년 3월부터 2013년 12월까지 코스닥 시장에 신규 상장한 모든 기업을 대상으로 일반투자자의 배정비율과 저평가율 간의 관계에 대해 실증분석 한다. 본 연구의 결과는 다음과 같이 요약될 수 있다. 일반투자자 배정비율의 20% 초과 여부가 저평가율에 있어 유의한 설명력을 갖는지 회귀식을 통해 검증한 결과 1% 유의수준 하에서 의미가 있다고 나타난다. 나아가 일반투자자의 배정비율도 5% 유의수준 하에서 저평가율에 대해 설명력을 갖는 것으로 나타난다. 저평가에 대한 가설은 여러 가지가 존재하지만, 일반투자자 배정비율이 높으면서 저평가가 뚜렷이 나타난다면, 상장 후 주가 및 실적에 대한 공모회사의 자신감을 볼 수 있는 간접적인 신호(Signalling)라고 볼 수 있다. 결론적으로 본 연구는 코스닥 IPO시장에서 공모주식 청약 시 일반투자자의 배정비율이 저평가 정도에 유의한 영향을 미치는 주요한 설명변수로 이용될 수 있음을 실증분석을 통해 확인한다.

• 대한간호학회지
• /
• 제24권2호
• /
• pp.278-302
• /
• 1994

This article reviewed and analyzed 39 studies on self efficacy theory applied to health related behavior. The following analysis was done : 1) study subjects 2) measurement tools 3) analysis according to the type of research design(intervention re-search, explanatory research). Some findings are summarized as follows : The study subjects were both healthy people in various developmental stages and patients with various illnesses. The health related behaviors examined in the studies were also various including exercise, smoking cessation, self care behaviors, etc. The measurement of self efficacy was done with specific tools in most studies. In the tools, activities that measured the health behavior domain were listed according to increasing difficulty or contexual arrangement or in combination of both of them The analysis of 17 intervention research studies showed that generally the intervention program increased the self efficacy level of subjects and then the increased strongly self efficacy influenced behavioral changes. Most studies used more than one intervention method for increasing the self efficacy level. these were derived from sources of self efficacy suggested by Bandura. The analysis of 21 explanatory research studies showed that self efficacy strongly influenced behavior change and persistence. The major independent variable to affect the self efficacy was performance accomplishment in the past. Self efficacy explained more of the variance in health related behavior when it was applied with the variables in the health belief model, health promotion model, and reasoned action theory. On the basis of the above findings, the following suggestions are made : 1. For a desirable research design, self efficacy should be the intervening variable. That is, desirable designs would include intervention-self efficacy-behavior in intervention research studies and antecedent-self efficacy-behavior in explanatory research studies. 2. More prospective, longitudinal studies are needed to test the effect of self efficacy on persistence in health related behavior. 3. Studies comparing the effects of intervention methods are needed for each health related behavior, subject group, and context. 4. It is necessary to develop a reliable, valid measurement tool for self efficacy for each health related behavior. 5. Studies to differenciate the effect of self efficacy from that of outcome expectation on the health related behavior are necessary. 6. The antecedents of self efficacy should be investigated further.

• 한국축산식품학회지
• /
• 제43권1호
• /
• pp.113-123
• /
• 2023

This study investigated the predictability of cooking loss of pork loin through relatively easy and quick measurable quality properties. The pH, color, moisture, protein content, and cooking loss of 100 pork loins were measured. The explanatory variables included in all linear regression models with an adjust-r² value of ≥0.5 were pH and the protein content. In the linear regression model predicting cooking loss, the highest adjust-r² value was 0.7, with pH, CIE L^*, CIE b^*, moisture, and protein content as the explanatory variables. In 30 pork loins, electrical conductivity was additionally measured, and as a result of linear regression analysis for predicting cooking loss, the highest adjust-r² value was 0.646 with electrical conductivity measured at 40 Hz, with pH and color as the explanatory variables. Ordinal logistic regression analysis was performed to predict the three grades (low, middle, and high) of loin cooking loss using pH, color, and 40 Hz electrical conductivity as the explanatory variables, and the percent concordance was 93.8%. In conclusion, the addition of electrical conductivity as an explanatory variable did not increase the prediction accuracy of the linear regression model for predicting cooking loss; however, it was demonstrated that it is possible to predict and classify the cooking loss grade of pork loin through quality properties that can be measured quickly and easily.

• 유통과학연구
• /
• 제13권9호
• /
• pp.47-57
• /
• 2015

Purpose - The aim of this paper is to investigate the explanatory power of the Pecking-order theory (the cost of financing increases with asymmetric information) among Korean retail firms from the perspective of debt capacity. According to the Pecking-order theory, a firm's first preference is to use internal funds for its capital needs, its next preference is the issuance of debt, and its last preference is the issuance of equity; this is due to the information asymmetry problem between existing shareholders and investors. However, prior empirical studies, such as Lemmon and Zender (2010), argue that the entire sample test for the Pecking-order theory could be misleading due to the different levels of debt issuance capability of each of the individual firms; in fact, they confirm that the explanatory power of the Pecking-order theory improves after taking into account the differences in debt capacity of the U.S. firms they examined. This paper implements a case study approach among Korean retail firms to examine the relationship between debt capacity and the explanatory power of the Pecking-order theory in Korea. Research design, data, and methodology - This study uses the sample of public retail firms on the Korea Composite Stock Price Index (KOSPI) from the time period of 1990 to 2013. We gather related financial and accounting statements from the financial information firm WISEfn. Credit rating information is provided by the Korea Investor Service. We employ the models of Lemmon and Zender (2010) and Son and Kim (2013) to measure a firm's debt capacity. Their logit models use the rating dummy variable as a dependent variable and incorporate other firm characteristics as independent variables to estimate debt capacity. To test the Pecking-order theory, we adopt variants of the financing deficit model of Shyam-Sunder and Myers (1999). In the test of the Pecking-order theory, we consider all of the changes in total debt obligations, current debt obligations, and long-term debt obligations. Results - Our main contribution to the literature is our confirmation of the predicted relationship between debt capacity and the explanatory power of the Pecking-order theory among Korean retail firms. The coefficients on financing deficits become greater as a firm's debt capacity improves. This is consistent with the results of Lemmon and Zender (2010). The coefficients on the square of the financing deficits are also negative for the firms in the largest debt capacity group, which is also consistent with the predictions in prior literature. Conclusions - This study takes a case study approach by examining Korean retail firms. We confirm that the Pecking-order theory explains the capital structure of retail firms more appropriately, after taking into account the debt capacity of each firm. This result suggests the importance of debt capacity consideration in the testing of the Pecking-order theory. Our result also implies that there has been a potential underestimation of the explanatory power of the Pecking-order theory in existing studies.

• 디지털융복합연구
• /
• 제11권4호
• /
• pp.351-359
• /
• 2013

본 연구는 사회인구학적 특성, 퇴원특성, 의료기관특성이 정신분열병 환자의 재원일수에 영향을 미치는 요인을 규명하고자 시행하였다. 2004~2008년의 퇴원심층조사 자료 중에서 주진단이 정신분열병인 2,239명의 환자가 최종 연구대상이었다. SPSS 18.0 프로그램을 이용하여 사회인구학적 특성, 퇴원특성, 의료기관 특성을 설명변수로 하고, 재원일수를 종속변수로 구분하여 설명변수를 순차적으로 투입한 위계적 회귀분석을 실시하였다. 분석결과 사회인구학적 특성이 가장 큰 설명력을 나타냈으며, 퇴원특성보다는 병원특성 설명변수를 추가할 경우 설명력의 변화가 크게 나타났다. 재원일수에 주요하게 영향을 미치는 요인은 남자, 의료급여1종, 충청도, 병상규모 등으로 밝혀졌다. 본 연구는 이차 자료를 이용하였기 때문에 정신분열병 환자의 재원일수를 보다 잘 설명할 수 있는 추가 변수의 제한이 있었다. 그럼에도 불구하고 국가적 차원의 대규모 데이터를 분석하여 규명하였다는데 의의가 있다. 재원일수의 감축 노력은 환자 개인특성뿐 아니라 병원특성에 더욱 중점을 두고 국가적 노력이 필요함을 제안한다.

• 응용통계연구
• /
• 제8권1호
• /
• pp.61-74
• /
• 1995

오늘날에는 컴퓨터를 이용한 다양한 그래프기법의 개발로 자료로부터 정보를 직접적으로 얻는 것이 용이하다. 특히 최근에 발표된 R-코드(Cook과 Weisberg, 1994)는 다양한 2차원, 3차원 플롯 뿐만 아니라 축의 회전과 여러가지 모형에 대한 적합성을 제시하므로 보다 쉽게 자료에 적합한 모형을 시각적으로 분석할 수 있게 하였다. 그러나 그래프는 3차원 이상의 공간을 표현할 수 없기 때문에 하나의 반응변수와 세개이상의 설명변수 사이의 관계를 직접적으로 표현하는 것이 불가능하다. 이와 관련하여 Li(1991, 1992)에 의하여 제시된 SIR, pHd 방법과 Cook과 Weisberg(1991)에 의하여 제시된 SAVE는 설명변수들의 선형결합을 이용하여 효과적으로 설명변수들의 차원을 줄이는 방법을 제시하였다. 본 연구에서는 Li에 의하여 제시된 pHd 방법을 반응변수가 2개이상인 다변량 반응변수 모형에 적용하는 방법을 연구하였다. pHd 방법의 적용에는 많은 계산과정이 요구되는데, 이러한 계산과 다양한 플롯은 R-코드를 이용하였다.

• 대한가정학회지
• /
• 제46권8호
• /
• pp.121-133
• /
• 2008

The purpose of this study was to decipher the BRQ (Brand Relationship Quality) Construct of fashion product consumers and to identify influencing variables. Consumption values of consumers, fashion product attributes and marketing communication instruments were investigated as affecting variables. A questionnaire was distributed to 350 women between the age of 20 and 50 living in Seoul and other metropolitan areas from Mar. 5 to Mar. 18, 2008. Collected data were analyzed by factor analysis, reliability, multiple regression, frequency and percentage using SPSS (Version 15,0) statistical Package. The results were as follows: Firstly, the BRQ Construct of fashion product consumers was proved to be composed of five factors; 'self-connective love', 'pursuing symbol', 'trust', 'intimacy', and 'knowledge'. Secondly, consumption values had a significant effect on the BRQ Construct and their explanatory power was 32%, 'Epistemic value' was the most significant variable affecting BRQ. 'Functional value' had a significant influence on some BRQ factors. Thirdly, explanatory power of fashion product attributes affecting BRQ was 45%. 'Aesthetic attribute' had significant effect on all BRQ factors, while 'functional attribute' showed significant effect on only 'pursuing symbol', 'trust', and 'intimacy'. Finally, marketing communication mixes showed 21% explanatory power and 'word of mouth' and 'display' among marketing communication mixes were found as the most influential variables affecting the BRQ Construct.

• Journal of the Korean Data and Information Science Society
• /
• 제23권3호
• /
• pp.535-542
• /
• 2012

소비자물가지수는 국가의 중요한 경제 척도 중의 하나이다. 본 연구에서는 4개 도시, 서울, 부산, 대구, 광주지역의 소비자물가지수를 연구하였다. 자료는 모두 통계청에서 발췌하였고, 기간은 1998년-2011년 월별자료이며, 시계열분석 기법인 자기회귀오차모형으로 분석하였다. 소비자물가 분석을 위한 설명변수는 9가지 경제변수인 경기동행지수, 미국환욜, 생산자물가지수, 원유수입단가, 원유수입물량, 국제경상수지, 수입물가지수, 실업율, 화폐통화량을 사용하였다. 분석 결과, 자기회귀오차모형으로 각 지역별 소비자물가지수를 46%-52% 정도 설명할 수 있다.

• 가족자원경영과 정책
• /
• 제6권1호
• /
• pp.103-115
• /
• 2002

The purpose of this study was to analyze which factor, influenced the business and family performance success. Data were obtained from 248 family households. Proxy variable of the business performance was gross business income and of the family performance was the Family AFGAR scores. The multiple regression analysis was conducted for both the business performance equation and family performance equation. The main results of this study were as followings: The results indicated the effects of various business and family characteristics on performance and their contributions to the business and family performance model. Nine explanatory variables such as sex, being home-based, number of hours worked per week, number of family employee, number of nonfamily employee, total asset, the presence of young child under 6, nonbusiness income, and role conflicts were statistically significant in the business performance equation and three explanatory variables such as the hours worked per week, family stress scores, and role conflicts were statistically significant in the family performance equation. The results indicated the need for a more comprehensive view of family business performance.