지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학위논문
- 저자정보
- 지도교수
- 이화운
- 발행연도
- 2020
- 저작권
- 부산대학교 논문은 저작권에 의해 보호받습니다.
이용수40
이 논문의 연구 히스토리 (2)
- 이 논문의 후속연구가 궁금하신가요?
- 연관 학술논문 또는 학술발표를 통해 보다 발전된 연구결과를 확인하실 수 있습니다.
- 이 논문의 연구 히스토리 확인하기
초록· 키워드
오류제보하기
본 연구는 지속적인 노령화가 이루어지고 있는 부산지역을 대상으로 지역적 환경 보건정책의 의사결정에 도움을 주고자 건강취약집단을 미성년자집단(DayPA_U20)과 65세 이상의 노령층집단(DayPA_O65)으로 구분하고 이를 중심으로 기상 및 대기오염농도가 호흡기 외래 환자 수(DayPA_A)에 어떤 영향을 미치는지 통계적으로 분석하고자 하였다.
먼저, 부산지역의 2009년 1월 1일에서 2018년 12월 31일까지 10년 자료를 이용하여 호흡기 외래 환자 수와 기상 및 대기오염농도와의 상관관계를 분석하여 상호관련성을 파악하였으며 자료의 시계열적 특성을 고려하여 기상 및 대기오염농도의 시계열 성분 분리를 통해 상관 분석을 실시하였다. 그 결과 풍속을 제외한 기상요소의 상관계수가 대기오염농도의 상관계수보다 상대적으로 높은 것으로 나타났다. 대기오염농도가 두 건강취약집단에 미치는 영향을 비교 분석하였을 경우 근소한 값이긴 하나 DayPA_U20과 더 큰 상관성을 보여주었고, 시계열 성분 분리 후의 계절적 변동성분과의 상관 분석은 DayPA_O65에서 더 큰 상관을 보여주었다.
자동회귀누적이동평균(Autoregressive integrated moving average, ARIMA)를 이용한 시계열 분석 결과 호흡기 외래 환자 수는 7일 주기의 뚜렷한 사회적 주기를 가지는 것으로 나타났으며, 기상요소 및 대기오염농도는 4계절 또는 1년의 장주기를 따르는 자료의 변동을 보여주었다.
일반화부가모형(Generalized additive model, GAM)을 이용한 개별 기상요소의 상대 위험도를 분석한 결과 DayPA_O65의 경우는 DayPA_A의 기상요소별 상대 위험도와 비슷하게 나타나며 기온, 풍속, 이슬점 온도, 상대습도, 수증기압이 낮아질수록 호흡기 외래 환자가 유의미하게 증가하는 것으로 나타났다.
대기오염농도에 대한 상대 위험도 분석을 위한 통계 모델은 일평균 기온-대기압을 상호 작용 항으로 하는 배경을 이용하여 개별오염물질들에 대한 상대 위험도를 분석하였다. CO, NO2, PM10의 증가는 DayPA_A 및 DayPA_O65를 유의미하게 증가시키는 것으로 나타났다. 일평균 PM10이 22.6 ㎍/㎥ 증가 시 DayPA_A는 1.29%, DayPA_O65은 2.45%의 증가율을 보인다. 일평균 NO2가 0.008 ppm 증가할 경우 DayPA_A은 3.41% 증가를, DayPA_O65은 4.33%의 증가율을 보인다.
복합오염물질 모델에 대한 상대 위험도 결과를 보면 일평균 PM10이 22.6 ㎍/㎥ 증가 시 DayPA_A는 1.58%, DayPA_U20은 1.24%, DayPA_O65는 2.01%의 증가율을 보인다.
ARIMAX 모델과 GAM을 이용한 환자 수 예측 결과는 환자군별로 다른 특성을 보였다. DayPA_A, DayPA_U20의 경우는 ARIMAX의 IOA 및 에러 지표에서 우수한 결과를 보였으며, DayPA_O65에서는 GAM NO2 모델이 가장 좋은 결과를 보여주었다.
ARIMAX 모델은 시계열 데이터의 동특성을 반영하여 예측을 수행하기 때문에 GAM 통계 모델보다는 반응성이 뛰어나 급격한 수치의 변동에서도 적절한 예측이 가능했으나 여름철 부근에서는 오히려 과대추정의 영향이 나타나 음의 환자 수를 예측하는 오류도 나타났다. GAM은 통계 모델의 특성상, 이상치(outlier)으로 분류될 수 있는 급격한 환자 수의 변동을 설명하지는 못하였으나 장주기의 추세를 잘 예측하는 결과를 보여주었다.
먼저, 부산지역의 2009년 1월 1일에서 2018년 12월 31일까지 10년 자료를 이용하여 호흡기 외래 환자 수와 기상 및 대기오염농도와의 상관관계를 분석하여 상호관련성을 파악하였으며 자료의 시계열적 특성을 고려하여 기상 및 대기오염농도의 시계열 성분 분리를 통해 상관 분석을 실시하였다. 그 결과 풍속을 제외한 기상요소의 상관계수가 대기오염농도의 상관계수보다 상대적으로 높은 것으로 나타났다. 대기오염농도가 두 건강취약집단에 미치는 영향을 비교 분석하였을 경우 근소한 값이긴 하나 DayPA_U20과 더 큰 상관성을 보여주었고, 시계열 성분 분리 후의 계절적 변동성분과의 상관 분석은 DayPA_O65에서 더 큰 상관을 보여주었다.
자동회귀누적이동평균(Autoregressive integrated moving average, ARIMA)를 이용한 시계열 분석 결과 호흡기 외래 환자 수는 7일 주기의 뚜렷한 사회적 주기를 가지는 것으로 나타났으며, 기상요소 및 대기오염농도는 4계절 또는 1년의 장주기를 따르는 자료의 변동을 보여주었다.
일반화부가모형(Generalized additive model, GAM)을 이용한 개별 기상요소의 상대 위험도를 분석한 결과 DayPA_O65의 경우는 DayPA_A의 기상요소별 상대 위험도와 비슷하게 나타나며 기온, 풍속, 이슬점 온도, 상대습도, 수증기압이 낮아질수록 호흡기 외래 환자가 유의미하게 증가하는 것으로 나타났다.
대기오염농도에 대한 상대 위험도 분석을 위한 통계 모델은 일평균 기온-대기압을 상호 작용 항으로 하는 배경을 이용하여 개별오염물질들에 대한 상대 위험도를 분석하였다. CO, NO2, PM10의 증가는 DayPA_A 및 DayPA_O65를 유의미하게 증가시키는 것으로 나타났다. 일평균 PM10이 22.6 ㎍/㎥ 증가 시 DayPA_A는 1.29%, DayPA_O65은 2.45%의 증가율을 보인다. 일평균 NO2가 0.008 ppm 증가할 경우 DayPA_A은 3.41% 증가를, DayPA_O65은 4.33%의 증가율을 보인다.
복합오염물질 모델에 대한 상대 위험도 결과를 보면 일평균 PM10이 22.6 ㎍/㎥ 증가 시 DayPA_A는 1.58%, DayPA_U20은 1.24%, DayPA_O65는 2.01%의 증가율을 보인다.
ARIMAX 모델과 GAM을 이용한 환자 수 예측 결과는 환자군별로 다른 특성을 보였다. DayPA_A, DayPA_U20의 경우는 ARIMAX의 IOA 및 에러 지표에서 우수한 결과를 보였으며, DayPA_O65에서는 GAM NO2 모델이 가장 좋은 결과를 보여주었다.
ARIMAX 모델은 시계열 데이터의 동특성을 반영하여 예측을 수행하기 때문에 GAM 통계 모델보다는 반응성이 뛰어나 급격한 수치의 변동에서도 적절한 예측이 가능했으나 여름철 부근에서는 오히려 과대추정의 영향이 나타나 음의 환자 수를 예측하는 오류도 나타났다. GAM은 통계 모델의 특성상, 이상치(outlier)으로 분류될 수 있는 급격한 환자 수의 변동을 설명하지는 못하였으나 장주기의 추세를 잘 예측하는 결과를 보여주었다.
목차
I. 서론 ··················································································································································· 11. 연구 배경과 목적 ···························································································································· 1가. 연구 배경 ······································································································································ 1나. 연구 목적 ······································································································································ 3II. 연구 방법 ············································································································································ 41. 연구대상 지역 및 연구자료 ·········································································································· 4가. 연구대상 지역 ······························································································································ 4나. 기상자료 및 대기오염농도자료 ································································································ 4다. 호흡기 외래 환자 수 ·················································································································· 62. 통계 분석 방법 ································································································································ 8가. 상관 분석 ······································································································································ 8나. 자동회귀누적이동평균모델(Autoregressive integrated moving average, ARIMA) ······· 9다. 일반화가법모형(Generalized additive model, GAM) ························································· 13라. 시계열 자료의 예측 오차와 평가 ·························································································· 16III. 통계 분석 결과 ······························································································································· 181. 연구자료의 특성 및 분석 ············································································································ 18가. 전체적인 연구자료의 범위 ······································································································ 18나. 연구자료의 시계열적 특성 ······································································································ 202. 기상요소 및 대기오염농도와 호흡기 외래 환자 수와의 상관관계 ···································· 33가. 상관계수의 유의성 ···················································································································· 33나. 대기오염농도와 호흡기 외래 환자 수와의 상관관계 ························································ 34다. 기상과 호흡기 외래 환자 수와의 상관관계 ········································································ 37라. 상관관계 분석 결과 ·················································································································· 393. ARIMA를 이용한 시계열 분석 ··································································································· 40가. 호흡기 질환자 수 시계열 분석 ······························································································ 41나. 대기오염농도 시계열 분석 ······································································································ 45다. 기상자료 시계열 분석 ·············································································································· 514. GAM을 이용한 상대 위험도 분석 ····························································································· 55가. 기상요소의 상대 위험도 분석 ································································································ 55나. 대기오염농도의 상대 위험도 분석 ························································································ 58IV. 예측 모델 ········································································································································ 771. ARIMAX 시계열 예측 ··················································································································· 782. GAM 시계열 예측 ························································································································· 853. 예측 결과 비교 ······························································································································ 94V. 결론 ··················································································································································· 98참고문헌 ··············································································································································· 100Abstract ·············································································································································· 103
최근 본 자료
댓글(0)
0