• 대한산업공학회지
• /
• 제42권5호
• /
• pp.314-326
• /
• 2016

The purpose of variable selection techniques is to select a subset of relevant variables for a particular learning algorithm in order to improve the accuracy of prediction model and improve the efficiency of the model. We conduct an empirical analysis to evaluate and compare seven well-known variable selection techniques for multiple linear regression model, which is one of the most commonly used regression model in practice. The variable selection techniques we apply are forward selection, backward elimination, stepwise selection, genetic algorithm (GA), ridge regression, lasso (Least Absolute Shrinkage and Selection Operator) and elastic net. Based on the experiment with 49 regression data sets, it is found that GA resulted in the lowest error rates while lasso most significantly reduces the number of variables. In terms of computational efficiency, forward/backward elimination and lasso requires less time than the other techniques.

• 한국작물학회지
• /
• 제51권4호
• /
• pp.269-273
• /
• 2006

Multiple linear regression analysis was conducted to interpretate the relationship between sesame grain yield and its components under early sowing cropping condition. The t test showed that stem length, number of capsules per plant, 1000 seeds weight and seed weight per plant gave significant contribution to sesame grain yield, therefore those variables were assumed to mostly influenced components to grain yield of sesame. In the stepwise regression analysis, the predicted equation for sesame grain yield per square meter (Y) was Y = -7.900 + 0.150X1 + 0.461X5 + 15.553X6 + 8.543X7. Meanwhile, F value showed that stem length, number of capsules per plant and seed weight per plant gave significant contribution to sesame grain yield, while 1000 seeds weight did not significantly show. Based on the results, it is reasonable to assume that high yield. potential of sesame under early sowing cropping condition would be obtained by selecting breeding lines with long stem length, number of capsules per plant, and seed weight per plant, which was different result at the late sowing cropping condition in which days to flowering and maturity were assumed to be more affected factors to the sesame grain yield.

• 가정과삶의질연구
• /
• 제23권1호
• /
• pp.19-29
• /
• 2005

The purpose of this study was to investigate ego-resilience and social adaptability of the children and adolescents in institutional care. The study also investigated the protective factors and the risk factors on ego-resilience and social adaptability and their relationship. The subjects in this study were 140 children and adolescents in institutional care. The research data were analyzed using Frequency, Pearson's Correlation, Stepwise Multiple Regression Analysis and T-Test in SPSS WIN 10.0 program. The major findings were as follows. First, ego-resilience of the children and adolescents in institutional care was related to locus of control, depression and anxiety and unsatisfaction of institutional care. Result of stepwise multiple regression analysis showed that depression and anxiety and locus of control in the order had significant effects on ego-resilience. Second, social adaptability of the children and adolescents in institutional care was related to social support and parent-child bond. Unsatisfaction of institutional care and social support in the order had significant effects on social adaptability by stepwise multiple regression analysis. Third, the children and adolescents who had higher level of ego-resilience reported better social adaptability. The overall research results revealed that the risk factors such as unsatisfaction of institutional care and depression and anxiety than the protective factors had more significant effects on ego-resilience and social adaptability of children and adolescents in institutional care. This finding suggested that elimination of risk factors rather than development of protective factors was a more important task for the welfare of children and adolescents in institutional care.

• 한국환경보건학회지
• /
• 제7권1호
• /
• pp.9-20
• /
• 1981

This study was performed to determine the correlationship between temperature and overall removals of BOD, SS and to demonstrate the effect of temperature on treatment performance. These data for a period from February 1, 1977 to January 31, 1980 were obtained from the Cheong-Gye Cheon Sewage Treatment plant. The results of correlation and stepwise multiple regression analysis were as follows. 1) Secondary effluent BOD and SS showed negative correlationship with water temperature, with correlation coefficient of -0.1710, and -0.1654 respectively. 2) Correlation coefficient of BOD, SS removal rate and water temperature were 0.1823 and 0.0429 respectively. 3) Regresion equation for estimate of BOD removal rate was as follows $\widehat{Y}_1$ (BOD removal rate)=63.9994+0.5442X(water temperature). And BOD removal rate showed non significant change according to the water temperature. 4) Regression equation for estimate of SS removal rate was as follows $\widehat{Y}_2$ (SS removal rate)=61.6881+0.1514X(Water temperature). And SS removal rate showed non significant change according to the water temperature. 5) According to the Stepwise Multiple Regression analysis, water temperature ranked second order in the BOD removal rate estimation and the equation was as follows $\widehat{Y}_1$ (BOD removal rate)=69.7398+0.2665 $X_1$ (Primary effluent BOD)+0.3562 $X_2$ (Water temperature)-0.0122 $X_3(Flow)+4413.271X_4$ (Organic Loading).

• 대한토목학회논문집
• /
• 제26권1B호
• /
• pp.39-50
• /
• 2006

홍수를 예측하기 위해서 국내 5대강 유역의 홍수통제소는 저류함수모형을 사용하고 있으며 현재까지 홍수예측에 대한 많은 연구가 이루어지고 있다. 이에 본 논문에서는 현재 홍수통제소에서 사용되고 있는 저류함수모형과 과거의 강우-수위 관계를 이용한 회귀분석(regression analysis), 그리고 인공신경망(artificial neural network)을 이용하여 홍수를 예측하고 이를 비교, 분석하고자 하였다. 저류함수모형의 경우는 홍수통제소의 대표매개변수와 보정된 최적(평균)매개변수를 적용하였다. 그리고 회귀분석과 인공신경망은 1995~2001년까지의 홍수사상 중 4개의 홍수사상을 선택하여 회귀계수를 구하고 역전파(backpropagation) 알고리즘을 사용하여 학습을 시켰다. 그 결과 저류함수모형의 경우 최적 매개변수를 이용하였을 때 기존의 홍수통제소에서 사용하고 있는 대표매개변수보다 예측이 개선되었으며, 회귀분석의 방법인 다중회귀분석, Robust 회귀분석, Stepwise 회귀분석을 이용한 홍수예측은 비교적 정확한 결과를 얻을 수 있었다. 역전파 알고리즘을 사용한 인공신경망의 경우도 회귀분석을 이용한 홍수예측보다는 다소 못하였지만 정확한 결과를 얻을 수 있었다.

• 대한가정학회지
• /
• 제45권4호
• /
• pp.29-42
• /
• 2007

The purpose of this study is to investigate the leisure and social support types which affect elderly depression. The study sample comprised 275 cases, and the analysis was performed by t-test, ANOVA, multiple regression and stepwise regression on SPSS ver. 10.0. The following three principle results were obtained: First, the types of leisure and social support differed according to domographic profiles. Second, among 6 sub-factors of leisure type, sports and viewing impressions activities contributed positively to elderly depression, as did emotional and appraisement supports among 4 sub-factors of social support. Third, stepwise regression analysis, conducted to determine the effect weights of factors of leisure types and social supports, showed that emotional support, social relationships activity, and viewing impressions activity strongly affected elderly depression in the order listed. Based of these results, the study suggested methods for developing an active leisure activities program which is necessary to minimize elderly depression.

• 대한간호학회지
• /
• 제33권6호
• /
• pp.762-771
• /
• 2003

Purpose: This study was undertaken in order to determine factors affecting health promoting lifestyle of elderly women. Method: The subjects were 299 elderly community residing women over the age 65 living in 2 large cities. The instruments used for this study were a survey of general characteristics, health promoting lifestyle(47items), perceived health status(6 items), self-efficacy(l7 items), and social support(18 items). The data was analyzed using descriptive statistics, t-test, ANOVA, Pearson correlation coefficients, and Stepwise Multiple Regression. Result: The result of the study are as follows: The average item score for the health promoting lifestyle was 2.46, the highest score on the subscale was interpersonal support(M=2.83). A significant difference between age, education level, income, experience of smoking, alcohol, exercise, and health promoting lifestyle were found. Stepwise multiple regression analysis revealed that the most powerful predictor of health promoting lifestyle was social support(55%). Conclusion: Social support accounted for 54% of the variance in health promoting lifestyle in the elderly women. Therefore, health promoting programs that increase social support should be developed to promote a healthy lifestyle of the elderly women.

• 간호행정학회지
• /
• 제12권4호
• /
• pp.597-603
• /
• 2006

Purpose: This study was to identify effect of nursing organizational culture on job satisfaction of preceptors. Method: The subjects were 73 preceptors who were working for the 4 of university hospitals in Seoul, Daejeon and Kyong Ki. This study was conducted from Aug to Oct 2005. The data was collected by questionnaires and were analyzed using descriptive statistics, pearson correlation coefficient, stepwise multiple regression. Result: Job satisfaction of preceptors were significantly correlated among nursing organizational culture. The stepwise multiple regression analysis on job satisfaction revealed that nursing organizational culture was a powerful predictor of job satisfaction. Conclusion: The findings indicate that nursing organizational culture is important for improving job satisfaction of preceptors. Therefore it is necessary to develop of good nursing organizational culture for successful preceptorship.

• 성인간호학회지
• /
• 제22권2호
• /
• pp.182-189
• /
• 2010

Purpose: The purpose of this study was to identify predictors on drinking problems of university students. Methods: The participants were 359 students by convenience sampling in P city. Data were collected from October 15 to December 20, 2007 using drinking problems questionnaire, Global assessment of recent stress scale, CES-D and self-esteem scale. The collected data were analyzed by descriptive statistics, Pearson's correlation coefficient, and stepwise multiple regression. Results: There were positive correlations between stress and drinking problems, between depression and drinking problems. Also, there were negative correlations between stress and self-esteem, between depression and self-esteem. In stepwise multiple regression analysis, the most powerful predictor of drinking problems was depression. Conclusion: The results suggest that effective management of depression and stress can be considered when developing a prevention program of drinking problems for university students.

• 성인간호학회지
• /
• 제19권5호
• /
• pp.37-46
• /
• 2007

Purpose: The purpose of this descriptive study was to identify predictors on health promoting behaviors of university students. Methods: The participants were 310 students by convenience sampling in G city. The data were collected through a questionnaire given to the students between September 10 to November 6, 2006. Descriptive statistics, t-test, ANOVA, Pearson's correlation coefficients, and stepwise multiple regression were used to analyze the data. Results: The mean score for the level of health promoting behaviors and optimism was 2.60(${\pm} .31$) and 2.70(${\pm} .55$) respectively. The most frequently used coping skill was the pursuit of social support. There were positive correlations between health promoting behaviors and optimism, between health promoting behaviors and active coping skill, and between health promoting behaviors and passive coping skill. In stepwise multiple regression analysis, the most powerful predictor of health promoting behaviors was active coping skill. Conclusion: The results suggest that coping skills with optimism can be considered when developing a health promotion program for university students.