오존농도의 예측 모형에 관한 연구 : 부산 신평동 지역

서의훈(Euy Hoon Suh),정연선(Yeon Seon Jeong) 한국자료분석학회 2001 Journal of the Korean Data Analysis Society Vol.3 No.3

본 연구에서는 부산 신평동 지역의 기후인자(풍속, 기온, 노점온도, 기압, 상대습도, 강우량, 일사량)와 대기오염물질(NO2, SO2, CO(단위: ppm), PM10(단위: ug/m3)) 자료를 사용하여, 1) 오존농도의 연도별, 월별, 시간별, 요일별 특성을 알아보고, 2) 오존농도와 기후인자, 대기오염물질 간의 관련성을 밝혔으며, 3) 오존고농도에 대한 1시간 후 예측모형을 연구하였다. This study was carried out to develop a statistical model for the surface ozone concentration. The one-hour average concentration of ozone(O3), nitrogen oxides(NO2), PM10, sulfur dioxide(SO2) and meteorological factors (i. e., wind velocity, temperature, dew point, pressure, relative humidity, precipitation, insolation intensity), during the period of 1998 to 2000, were analysed. First, we examined the patterns of the ozone concentration by monthly, hourly, annually and weekly. Secondly, in order to find out the degree of association among the ozone concentration, air pollutants(CO, NO2, PM10, SO2), and meteorological factors(wind velocity, temperature, dew point, pressure, relative humidity, precipitation, insolation intensity), we used the statistical methods of correlation analysis, analysis of variances. Finally, we examined the 1 hour prediction models of the ozone concentration using the Logistic regression analysis and decision tree analysis.

범죄율에 따른 도시의 군집화

서의훈(Euy Hoon Suh),양호숙(Ho Sook Yang) 한국자료분석학회 2000 Journal of the Korean Data Analysis Society Vol.2 No.3

이 논문은 범죄발생률과 지역 특성간의 관계를 파악하는 데 그 목적이 있다. 다섯 가지 범죄발생률에 따라 전체 도시를 분류하고 분류된 각각의 군집에 대한 지역적 특성을 살펴보았다. 범죄발생률이 다른 지역에 비해 매우 높은 동두천, 오산, 과천, 여천을 제외한 나머지 도시는 4개의 군집으로 분류되었다. 범죄발생률이 높은 지역은 이혼률이 높고 전화보급수와 유흥업소 수가 많으며 고령자비, 저소득자비, 학교수가 작은 것으로 나타났다. This study aims to explore the relationship between the regional characteristics and crime rates. We classify residential areas by the crime rates and find some distinguished characteristics in classified residential areas. All cities are classified into four clusters, except the four cities with very high crime rates(Dongduchun, Osan, Kwachun and Yeuchun). The cluster with high crime rates is characterized by high divorce rate and many pleasure resorts.

낙동강 부영양화와 수질환경요인의 통계적 분석

김미숙,정영륜,서의훈,송원섭,Kim, Mi-Suk,Chung, Young-Ryun,Suh, Euy-Hoon,Song, Won-Sup 한국조류학회(藻類) 2002 ALGAE Vol.17 No.2

Influences of vrious environmental factors on the eutrophication of Nakdong River were analyzed statistically using water samples collected from 1 January, 1999, to 30 September, 2001 at Namji area. The relationships between the concentration of chlorophyll α (eutrophication index) and environmental factors and were analyzed to develop a statistical model which can predict the status of eutrophication. The concentation of chlorophyll α ranged from 66.2 mg · $m^{-3}$ to 70.8 mg · $m^{-3}$ during dry winter season and the average concentration during this study period was 35.5 mg · $m^{-3}$ Namji area of Nakdong River was in the hypereutrohic stage in terms of water quality. Stephanodiscus sp. and Aulacoseria granulata var. angustissima were dominant species during the witnter to spring time and summer to autumn period, respectively. Based on the correlation analysis and the analysis of variance between chlorophyll α concentration and environmental factors, significantly high positive relationships were found in the order of BOD> pH> COD > KMnO₄ consumption > DO > conductivity > alkalinity. In contrast to these factors, significantly negrative relationships were found as in the order of $PO₄^{3-}-P$ >water level>the rate of Namgang-dam discharge > NH₃-N> the rate of Andong-dam discharge> the rate of Hapchoen-dam discharge. Based on the factors analysis of environmental factors on the concentration of chlorophyll α, we obtained five factors as follows. The first factor included water level, pH, turbiditiy, conductivity, alkalinity and the rate of Namgang-dam discharge. The second factor included water temperature DO, NH₄+-N, NO₃- -N. The third factor included KMnO₄ consumption COD and BOD. The fourth factor included the rate of Andong-dam discharge, the rate of Hapcheon-dam discharge, and the rate of Imha-dam discharge. The final factor included T-N T-P and $PO₄^{3-}-P$ > concentration. We derived two statistica models that can predict the occurrence of eutrophication based on the factors by factor analysis, using regression analysis. The first model is the stepwise regression model whose independent variables are the factors produced by factor analysis : chl α (mg · $m^{-3}$ = 42.923+(18.637 factor 3) + (-17.147 factor 1) + (-12.095 factor 5) + (-4.828 factor 4). The second model is the alternative stepwise regression model whose independent variables are the sums of the standardized main component variables:chl α (mg · $m^{-3}$ = 37.295+(7.326 Zfactor 3) + (-2.704 Zfactor 1)+(-2.341 Zfactor 5).

부산,경남 지역의 피부미용실 증가 추세에 대한 통계적 모형

하명희 ( Myung Hee Ha ),서의훈 ( Euy Hoon Suh ) 한국미용학회 2002 한국미용학회지 Vol.8 No.2

단체급식소 영양사의 종합적 품질경영(TQM) 수행과 조직문화와의 관계 규명에 관한 연구

조기원 ( Ki Won Cho ),서의훈 ( Euy Hoon Suh ),윤지영 ( Ji Young Yoon ) 한국식생활문화학회 2007 韓國食生活文化學會誌 Vol.22 No.2

논문에 대한 통계분석의 적정성에 관한 연구 : 경상대학교 석사학위 논문을 대상으로

김진희(Jin-Hee Kim),서의훈(Euy-Hoon Suh) 한국자료분석학회 1999 Journal of the Korean Data Analysis Society Vol.1 No.1

경험적 자료수집과 통계적 기법 적용에 대한 적정성을 평가하기 위해 일정한 점검표에 의거하여 1994년에서 1996년 사이에 경상대학교에서 발표된 석사학위 논문 중 통계적 기법을 적용한 논문을 대상으로 적정성 평가를 실시하였다. 적정성 평가를 통해 통계적 기법 적용에 대한 실태를 파악하고, 통계적 기법 적용에서의 오류를 방지하고자 하였다. We evaluated 141 master s degree theses, submitted to Gyeongsang National University during the period of 1994-1996, in terms of the adequacy of data collection and statistical analyses. The evaluation was conducted on the basis of check lists. We found several typical errors which was often made in the process of data collection and statistical analysis.

자전거 헬멧 착용의 연관 요인: 다수준분석

김장락(Jang-Rak Kim),서의훈(Euy-Hoon Suh) 한국보건교육건강증진학회 2020 보건교육건강증진학회지 Vol.37 No.3

Objectives: This study sought to identify factors associated with bicycle helmet use at both individual and regional levels. Methods: We analyzed data from community health surveys conducted in 2016 in 254 municipal districts (si, gun, and gu) by the Korea Center for Disease Control. In total, 49,105 bicycle riders were enrolled from among the 228,452 adults systematically sampled using the resident registration database. Regional (municipal)-level data were provided by K-Health ranking (Hallym University, South Korea), which represents a collection of indicators processed from various sources. Multilevel binary logistic regression analysis was performed for the dependent variable, a self-reported response of I always wear a helmet. Results: The percentage of bicycle helmet use was 13.7% among men and 6.2% among women. Regional-level factors accounted for 11% and 10% of the variance of bicycle helmet use among men and women, respectively. Among regional variables, the proportion (%) of the population that was elderly significantly decreased the odds of helmet use (adjusted odds ratio [AOR] = 0.97, 95% confidence interval [CI], 0.95-0.98 for men; AOR = 0.96, 95% CI, 0.94-0.98 for women). Conclusion: Community efforts to promote bicycle helmet use are needed, especially in areas with a large elderly population.

멀티미디어 통신을 위한 통계적 다중화기의 성능분석

안성진(Ahn SungJin),김수택(Kim SooTack),서의훈(Euy-Hoon Suh) 한국자료분석학회 2000 Journal of the Korean Data Analysis Society Vol.2 No.1

확률적으로 동일한 M개의 on-off VBR 원천들에서 발생하는 데이터를 하나의 ATM 링크 위로 다중화하는 통계적 다중화기를 분석하였다. 특히, 버스트 수준 다중화 전략과 셀 수준 다중화 전략을 비교하였다. 우리는 분석적으로 셀 도착에 대한 한 특정모형에 대해서 “ATM 서버 속도” 대 “고객 접근선속도” 비의 임계값이 존재한다는 것을 보였다. 모의실험 결과는 다중화기에 대한 분석으로부터 예상될 수 있었던 것을 확인시켜 주었다. We analysed a multiplexer supporting M statistically identical on-off sources on to an ATM link. In particular, we compared the performance of two multiplexing strategies: burst level multiplexig and cell level multiplexing. We showed analytically that, for a particular model for cell arrivals, there is a threshold of ATM trunk-speed to customer access-line-speed ratio below which burst level multiplexing is better for minimizing mean end-to-end burst delay. The simulation results confirmed what might be expected from the analytical modeling of the multiplexer.

진주시 피부관리 고객의 피부건강상태에 관한 통계적 연구

서의훈,하명희 한국미용학회 1999 한국미용학회지 Vol.5 No.2

In this paper, we sampled 260 clients who visited the skin care shops located in Chinju city during the period between July 26 and August 7, 1999. The aim of this study is, utilizing the survey data, (1) to investigate the socio-demographic characteristics(age, educational level, marital status, occupation, etc.) of the clients. (2) to make statistical tests about the effects of some factors on the facial health condition of women. The factors include: 1) childbirths, abortions, regularity of a menstrual period as physiological factors; 2) the degree of stress and life satisfaction in the family as psychological factors; 3) regularity of meal-taking, unbalanced diet, favorite foods, smoking and drinking as food-taking habits; 4) the frequency and duration of skin care.

부산 , 마산 , 진주지역 피부관리 고객의 사회인구학적 특성과 피부건강에 관한 비교 연구

서의훈,하명희 한국미용학회 2000 한국미용학회지 Vol.6 No.2

In this paper, we sampled 881 clients who visited the skin care shops located in Pusan, Masan and Chinju city during the period between July 26 and August 7, 1999. The aim of this study is, utilizing the survey data, to do a comparative study among three cities about (1) the socio-demographic characteristics(e.g., age, educational level, marital status, occupation, etc.) of the clients. (2) the effects of several factors on the facial health condition of women. The factors include: 1) childbirths, abortions, regularity of a menstrual period as physiological factors; 2) the degree of stress and life satisfaction in the family as psychological factors; 3) regularity of meal-taking, unbalanced diet, favorite foods, smoking and drinking as food-taking habits; 4) the frequency and duration of skin care.