• 한국산학기술학회논문지
• /
• 제20권12호
• /
• pp.652-662
• /
• 2019

본 연구의 목적은 청소년기의 총수면의 양과 그들의 정신적 건강 사이의 중요성을 알아보기 위해 2018년 청소년 건강행태온라인조사자료를 이용하여 총소년 총 수면시간과 자살생각 사이의 연관성을 규명하기 위해 수행하였다. 연구방법은 청소년건강온라인조사자료를 활용하여 카이제곱검정과 다중로지스틱회귀분석을 실시하여 결과를 도출하였다. 분석결과 수면의 질을 보정한 상태로 청소년들의 총 수면시간과 자살 생각 사이의 연관성을 분석한 결과 청소년들이 본인의 나이에 충분하다고 권장되는 시간을 충족하지 못하는 수면 양상을 보일 때 자살 생각을 하는 경향이 증가하는 것으로 나타났다. (OR: 1.17, 95% CI:1.06-1.29, p-vaule: 0.00) 이를 성별에 따라 층화분석한 결과 남학생의 경우에는 충분한 수면이 보장되면 자살 생각을 하는 빈도가 낮아지는 경향을 보이는 반면(OR:0.83, 95% CI:0.71-0.96, p-value:0.02), 여학생의 경우에는 남학생과는 반대로 충분한 수면이 보장되지 못할 때 자살 생각의 빈도가 높아지는 것으로 나타났다(OR:1.22, 95% CI:1.06-1.41, p-value:0.01). 이처럼 청소년들의 성별과 수면의 양상에 따라 총 수면시간이 자살 생각에 대해 미치는 영향이 다르기 때문에 10대의 자살 위험성을 낮추기 위해 이들의 생활습관에 맞는 새로운 정책적 고려가 필요할 것으로 생각된다.

• 한국산학기술학회논문지
• /
• 제21권4호
• /
• pp.145-155
• /
• 2020

Campbell(1990)이 수행의 다차원성을 제안한 이후 전통적 직무수행의 단일구조와는 달리 크게 과업수행, 맥락수행, 반생산적 업무행동으로 분류되어 왔다. 이에 본 연구의 목적은 (1) 지금까지 직무수행의 다양한 측면으로 제시된 과업수행, 맥락수행, 반생산적 업무행동으로 구성된 수행의 요인구조를 확인하고, (2) 이들 3요인을 기반으로 수행 유형을 도출한 후 (3) 수형 유형별 성격 변인들의 예측력을 비교하는 것이다. 이를 위해 국내 직장인 681명을 대상으로 자기보고식 온라인 설문조사를 실시하였다. 탐색적 및 확인적 요인분석 결과, 과업수행, 맥락수행, CWB로 이루어진 3요인 직무수행 모델이 확인되었다. 성격 변인과의 관련성도 각 수행 차원별로 구분되는 결과가 나타나 3요인 구조의 타당성이 입증되었다. 또한, 위 3요인을 기반으로 잠재프로파일 분석(Latent Profile Analysis, LPA)을 실시하여 4개 수행 유형을 도출하였다(즉, 모범적 수행자, 양심적 중간수행자, 반생산적 중간수행자, 양심적 저수행자). 다항 로지스틱 회귀분석 결과, 이들 4개 유형별로 각기 다른 성격 변인이 선행변인으로 작용하는 것으로 나타났다. 마지막으로 위 연구 결과를 토대로 본 연구의 시사점과 제한점에 대해 논의하였다.

• 응용통계연구
• /
• 제30권6호
• /
• pp.809-825
• /
• 2017

오디션, 육아, 버라이어티 등 다양한 예능 프로그램들의 수가 점점 증가하고 있다. 특히 종합편성채널이 개국한 이후에 예능 시장 경쟁이 심화되고 있다. 그에 따라 시청률과 회차에 대한 연구의 필요성이 대두되고 있다. 본 연구의 목적은 예능 프로그램 시청률과 회차의 예측모형을 제시하고 주요요인을 살펴보는 데 있다. 모형 적합 시 선형회귀모형, 로지스틱 회귀모형, LASSO 회귀모형, 랜덤 포레스트, 그래디언트 부스팅, 서포트 벡터 머신 등과 같은 다양한 분석 방법을 고려하였다. 예능 시청률 예측 모형에서는 첫 회가 방영되기 전과 방영된 후 두 가지 모형을 적합하였고, 회차 예측 모형에서는 예능 시청률 예측 모형의 예측치를 추가 변수로 생성하여 모형을 적합하였다. 그 결과 첫 회 방영 전 예능 시청률 예측에서는 방송사, 이전 시즌 시청률, 시작 연도, 기사 수가 큰 영향을 주는 것으로 나타났다. 첫 회 방영 후 예능 시청률 예측에서는 첫 회 시청률, 방송사, 예능 유형이 중요한 변수로 나타났으며, 두 모형 모두 랜덤 포레스트 모형에서 가장 좋은 결과를 보였다. 예능 회차 예측에서는 평균 시청률 예측치, 시작 연도, 예능유형, 방송국 등이 중요한 변수로 나타났다.

• 수산해양기술연구
• /
• 제30권3호
• /
• pp.172-181
• /
• 1994

The mesh selectivity of diamond and suare mesh cod-ends at the Southern Korean Sea and the East China Sea were compared for Pampus argenteus, Trachurus japonicus, Trichiurus lepturus. Selection trials were carried out using diamond and square mesh cod-end by trouser type cod-end with cover net. of which the mesh cod-end has four types : A(51.2mm), B(70.2mm), C(77.6mm), D(88.0mm). Selection curves and selection parameters were calculated using a logistic model. The results obained are summarized as follows : 1. Harvest fish : In B. C and D type selection range and fifty percent selection length of the square mesh were about 21mm, 11mm : 12mm, 18mm and 34mm, 5mm higher than those of the diamond mesh, respectively. Selection factor of master curve for the diamond mesh was 1.54 and for the square mesh was 1.68. The optimum mesh size for the diamond mesh was 97.4mm and for the square mesh was 89.3mm, the difference was 8.1mm. 2. Horse mackerel : In A type, selection range was nearly the same for the diamond and the square mesh, but fifty percent selection length of the square mesh was 43mm higher than the diamond mesh. In B. C and D type, selection range and fifty percent selection length of the square mesh were about 6mm, 3mm : 24mm, 21mm and 11mm, 42mm higher than those of the diamond mesh, respectively. Selection factor of master curve for the diamond mesh was 2.37, for the square mesh was 2.77. The optimum mesh size for the diamond mesh was 78.1mm and for the square mesh was 66.8mm, the difference was 11.3mm. 3. Hair tail : In A, B and C type, selection range of the square mesh was about 34mm, 8mm, 60mm higher than those of the diamond mesh. Fifty percent selection length for the diamond mesh was about 5mm, 7mm, 8mm higher than that of the square mesh. Selection factor of master curve for the diamond mesh was 3.11, for the square mesh was 3.48. The optimum mesh size for the diamond mesh was 64.3mm and for the square mesh was 57.5mm, the difference was 6.8mm.

• 수산해양기술연구
• /
• 제30권3호
• /
• pp.161-171
• /
• 1994

Mesh selection analysis for the trawl net were carried out at the Southern Korean Sea and the East China Sea during the 1992-1994 years by the training ship Seabada of the National Fisheries University of Pusan, using A type (51.2mm), B type (70.2mm), C type (77.6mm), D type (88.0mm) square mesh cod-ends. The fishing trials were made using bottom trawl of the trouser type cod-end with cover net. Selection curves and selection parameters were calculated by a logistic model for Sphyraena pinguis, Pampus argenteus, Trachurus japonicus, Pagrus major, Callanthias japonicus, Trichiurus lepturus. The results obtained are summarized as follows: 1. Red barracuda : Selection range and fifty percent selection length in the A type was 115.8mm, 292.8mm, respectively. 2. Harvest fish : Each selection range and fifty percent selection length in the B. C. and D type was 37.7mm, 113.8mm : 40.1mm, 131.7mm and 64.8mm, 148.6mm, respectively. Selection parameters of master curve were : slope, 3.81 : intercept, -6.4. Optimum mesh size was 89.3mm. 3. Horse mackerel : Each selection range and fifty percent selection length in the A, B, C and D type was 43.0mm, 148.3mm : 60.7mm, 183.2mm, 214.5mm and 91.4mm, 254.9mm, respectively. Selection parameters of master curve were : slope 2.30 : intercept, -6.4. Optimum mesh size was 66.8mm. 4. Red seabrem : Selection range and fifty percent selection length in the D type was 42.7mm, 203.4mm, respectively. 5. Yellowsail red bass : Selection range and fifty percent selection length in the A type was 84.0mm, 110.6mm, respectively. 6. Hair tail : Each selection range and fifty percent selection length in the A, B and C type was 59.7mm, 176.0mm : 100.9mm, 250.7mm and 178.6mm, 307.0mm, respectively. Selection parameters of master curve were : slope, 1.54 : intercept, -5.4. Optimum mesh size was 57.5mm.

• Journal of Preventive Medicine and Public Health
• /
• 제35권1호
• /
• pp.1-12
• /
• 2002

Objective : To estimate the socioeconomic costs of obesity in Korea,1998. Methods : The 1998 National Health and Nutrition Examination Survey(1998 NHNES) data was used and 10,880 persons who had taken health examinations were selected for study. Essential hypertension, NIDDM(non insulin-dependent diabetes mellitus), dyslipidemia, osteoarthritis, coronary heart disease, stroke were included as obesity related disease. The data of direct costs of obesity was obtained from the National Federation of Medical Insurance. The category of indirect costs was the loss of productivity caused by premature death and admission, time costs, traffic costs, nursing fees due to obesity. Multiple logistic regression model was developed to estimate prevalence odds ratio by obesity class adjusted demographic and socio-ecnomic factors and calculate PAF(Population Attributable Fraction) of obesity on obesity related disease. And we finally calculated the socioeconomic costs of obesity in relation to BMI with PAF. Results : The direct costs of obesity were 2,126 billion${\sim}965$ billion Won in considering out of pocket payment to uninsured services, and the indirect costs of obesity were 2,099 billion${\sim}1,086$ billion Won. Consequently, in considering out of pocket payment to uninsured services, the socioeconomic costs of obesity were 4.225 billion${\sim}2,050$ billion Won, which corresponded to about $0.094%{\sim}0.046%$ of GDP and $1.88%{\sim}0.91$ of total health care costs in Korea. Conclusions : Obesity represents a major health problem with significant economic implications for the society. This results are conservative estimates as far as all obesity related disease and all health care and indirect costs were not included due to missing information. further studies are needed to caculate socioeconomic costs of obesity more exactly.

• Journal of Preventive Medicine and Public Health
• /
• 제25권1호
• /
• pp.53-63
• /
• 1992

There are increasing interest and need for information on health care consumer with the significance of hospital marketing and strategic planning being increasingly emphasized. This study was conducted to investigate the criteria for selection of medical facilities according to the characteristics of health care consumer by the types of medical services on a sample of 1,500 population aged 20 years and above. Major findings are as follows ; 1. When considering the criteria for selection of medical facilities into two factors, namely, quality or convenience factors, convenience factor was the major contributor for outpatient and dental services whereas it was quality factor for inpatient services. 2. Females and those residing in large cities selected medical facilities based on convenience factor in the outpatient services. In the case of inpatient service, persons who considered their present health status to be good and whose ages were 50 years old and above choose medical facilities based on quality factor. 3. Persons who considered medical facilities to be profit-making tended to choose medical facilities based on convenience factor for outpatient services. There were no differences in the cases of inpatient and dental services. 4. There was no significant difference on the criteria for selection of medical facilities according to the decision maker for selection or trust on medical facilities. On the use of health service information, selection of medical facilities was based on quality factor for those who made more use of the information in the cases of outpatient and dental services. 5. Analysis using the logistic regression model on the criteria for the selection of medical facilities with the characteristics of health care consumer as independent variables was performed. The selection of medical facilities was significantly related with residential area, sex, and use of information on medical facilities for outpatient services and with age, average monthly income, and perception of health status for inpatient services. For dental services significant association with residential area and use of information on medical facilities was seen. The results of this study, despite some limitations, can be used as baseline data for marketing and strategic planning of hospital management.

• 한국수자원학회논문집
• /
• 제49권8호
• /
• pp.645-656
• /
• 2016

본 연구에서는 금강유역에 대한 지역홍수빈도분석을 실시하고 재현기간에 따른 홍수량을 추정하는 관계식을 제안하였다. 유역 내 유량자료의 수문학적 독립성과 동질성에 대한 검증을 위하여 Lag-1 자기상관성 분석, 동질성 검정, 이상치 검정, 불일치척도 검정을 수행하였다. 검정 결과, 금강유역의 대상 관측소들은 시간에 대하여 독립적이고 동질적 모집단에 속하며 이상치는 없었다. 일반 극치 분포(GEV), 3변수 대수정규 분포(LN-III), 피어슨-III 분포(P-III), 일반 로지스틱 분포(GLO), 일반 파레토 분포(GPA) 등 5개의 3변수 확률분포함수에 대한 L-모멘트비도와 평균가중거리(AWD), 그리고 $Z^{DIST}$ 적합도 산정 결과, GLO 분포함수가 금강유역의 최적 확률분포형으로 선정되었다. GLO 분포를 바탕으로 지역홍수빈도를 추정하는 회귀모형을 제안하였고, 강경 관측소의 관측 유량을 이용하여 회귀모형의 적용성을 검증하였다.

• 대한공간정보학회지
• /
• 제22권3호
• /
• pp.11-19
• /
• 2014

본 연구는 미래 기후변화 및 토지피복변화가 유역 내 유출량과 유사량의 거동에 가져올 영향을 분석하는데 목적이 있다. 기상자료는 IPCC 5차 평가보고서를 위해 새롭게 논의된 RCP 시나리오 중 서로 상반되는 4.5 및 8.5 시나리오의 기후전망 (2011~2100년) 이 사용되었으며, 토지피복지도는 RCP 4.5 및 8.5 시나리오의 사회 경제 스토리라인과 로지스틱 회귀모형 (LR)을 이용하여 개발 된 모델에 의해 구축되었다. 기후변화만 고려한 경우, 토지피복변화만 고려한 경우, 기후변화 및 토지피복변화 모두 고려한 경우의 세 가지 시나리오를 설정하고, 각 시나리오에 따른 유출량 및 유사량을 모의한 결과 계절적으로 매우 명확한 변화를 나타내었다. 기후변화는 봄과 겨울에 유출량을 증가시키고 여름과 가을에 유출량을 감소시키는 것으로 나타났으며, 유사량 역시 유출량과 동일한 변화의 양상을 보였다. 토지피복변화는 유출량을 증가시키는 반면 유사량은 감소시키며, 이는 도시화로 인한 불투수 면적의 증가에 의한 것으로 판단된다. 토지피복변화는 기후변화에 비해 유출량에 적은 영향을 끼치나, 기후변화에 의해 초래된 유출량 문제를 더욱 극대화 시킬 수 있다. 따라서 지속적인 수자원 관리를 위하여 기후변화의 잠재적 영향을 파악하고, 토지피복변화에 따른 적절한 수자원 대응 정책 마련이 필요할 것으로 판단된다.

• 보건행정학회지
• /
• 제21권2호
• /
• pp.329-348
• /
• 2011

This research analyzed 168,010 cases of death of the aged over 65 from 244,867cases of death excluding 7 unknown age cases from 244,874 all age cases of death by using the death data of the National Statistical Office for 2007 to figure out factors affecting the quality of causes of death statistics of the aged and to suggest the ways of improving the quality of death statistics of the aged in korea. This research tried to derive factors affecting ill-defined cause of death category in acordance with WHO's guidelines and to find causes of lowering the accuracy of causes of death statistics of the aged. This research identified the problems of causes of death statistics of the aged by using both demographic characteristics such as sex, age, marital status, educational attainment, residential region, region size and factors of death items as independent variable to find causes of ill-defined cause of death of the aged. Logistic regression analysis was executed to calculate the hazard ratio about the ill-defined causes of death of the aged and multiple regression analysis was conducted to derive factors affecting the ill-defined cause of death by regional groups through using these independent variables such as the component ratio of over age 65, female death rates, doctors insitutions rate, medical institutions rate, attaching rates of death certificate by neighborhood. As a results of this research, R-code was the highest of ill-defined causes of death, accounting for 82.1%, and senility death(R54) of R-code was the highest, accounting for 91.2%. through subdivided order distribution of the ill-defined causes of death of the aged. As ill-defined causes of death by regional groups, attaching rates of death certificate by neighborhood was the most important factor(p<0.05) and also showed regression model's description with 83.8% ($R^2$=83.8%). Furthermore, Jeon-nam was the highest in the regional groups and these regions such as Je-ju, Jeon-buk, Chung-nam were not only attaching the death certificate by neighborhood but also were high at the rate of ill-defined causes of death. Therefore, this research found that both reconsideration about death certificate by neighborhood and education for doctors who write death certificate were needed the most.