한국수자원학회논문집 (Journal of Korea Water Resources Association)

  • 제56권4호
  • /
  • Pages.261-272
  • /
  • 2023
  • /
  • 2799-8746(pISSN)
  • /
  • 2799-8754(eISSN)
한국수자원학회 (Korea Water Resources Association)
권호 표지
DOI QR코드

DOI QR Code

머신러닝 기법을 이용한 재해강도 분류모형 개발

Development of disaster severity classification model using machine learning technique

  • 이승민 (인하대학교 스마트시티공학과) ;
  • 백선욱 (인하대학교 스마트시티공학과) ;
  • 이준학 (인하대학교 스마트시티공학과) ;
  • 김경탁 (한국건설기술연구원 수자원하천연구본부) ;
  • 김수전 (인하대학교 사회인프라공학과) ;
  • 김형수 (인하대학교 사회인프라공학과)
  • Lee, Seungmin (Program in Smart City Engineering, Inha University) ;
  • Baek, Seonuk (Program in Smart City Engineering, Inha University) ;
  • Lee, Junhak (Program in Smart City Engineering, Inha University) ;
  • Kim, Kyungtak (Department of Hydro Science and Engineering Research, Korea Institute of Civil Engineering and Building Technology (KICT)) ;
  • Kim, Soojun (Department of Civil Engineering, Inha University) ;
  • Kim, Hung Soo (Department of Civil Engineering, Inha University)
  • 투고 : 2023.02.13
  • 심사 : 2023.03.24
  • 발행 : 2023.04.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

최근 급격한 도시화와 기후변화에 따라 재난에 의한 피해가 증가하고 있다. 국내 기상청에서는 표준 경보(주의보, 경보)를 전국적으로 통일된 표준 경보 기준(3시간 및 12시간 최대 누적강우량)에 따라 발령하여 재해에 따른 지역별, 재난 사상별 특성이 고려되지 않은 문제점이 있다. 따라서 본 연구에서는 서울특별시, 인천광역시, 경기도의 호우·태풍에 대한 재해 피해액 및 누적강우량을 활용하여 대상지역별 재해강도에 따른 단계별 기준을 설정하고, 강우에 따라 발생할 수 있는 재해의 강도를 분류하는 모형을 개발하고자 하였다. 즉, 본 연구에서는 호우·태풍에 의한 재해 피해액 누적 분포 함수의 분위별로 재해강도의 범주(관심, 주의, 경계, 심각 단계)를 분류하였고, 재해강도의 범주에 따른 누적강우량 기준을 대상 지자체별로 제시하였다. 그리고 지자체별 재해강도 분류모형 개발을 위해 4가지(의사결정나무, 서포트 벡터 머신, 랜덤 포레스트, XGBoost)의 머신러닝 모형을 활용하였는데 강우량, 누적강우량, 지속시간 최대 강우량(3시간, 12시간), 선행강우량을 독립변수로 이용하여 종속변수인 지자체별 재해강도를 분류하였다. 각 모형별 F1 점수를 이용한 정확도 평가 결과, 의사결정나무의 F1 점수가 0.56으로 가장 우수한 정확도를 보였다. 본 연구에서 제시한 머신러닝 기반 재해강도 분류모형을 활용하면 호우·태풍에 의한 재해에 대한 지자체별 위험 상태를 단계별로 파악할 수 있어, 재난 담당자들의 신속한 의사결정을 위한 기초 자료로 활용될 수 있을 것으로 판단된다.

In recent years, natural disasters such as heavy rainfall and typhoons have occurred more frequently, and their severity has increased due to climate change. The Korea Meteorological Administration (KMA) currently uses the same criteria for all regions in Korea for watch and warning based on the maximum cumulative rainfall with durations of 3-hour and 12-hour to reduce damage. However, KMA's criteria do not consider the regional characteristics of damages caused by heavy rainfall and typhoon events. In this regard, it is necessary to develop new criteria considering regional characteristics of damage and cumulative rainfalls in durations, establishing four stages: blue, yellow, orange, and red. A classification model, called DSCM (Disaster Severity Classification Model), for the four-stage disaster severity was developed using four machine learning models (Decision Tree, Support Vector Machine, Random Forest, and XGBoost). This study applied DSCM to local governments of Seoul, Incheon, and Gyeonggi Province province. To develop DSCM, we used data on rainfall, cumulative rainfall, maximum rainfalls for durations of 3-hour and 12-hour, and antecedent rainfall as independent variables, and a 4-class damage scale for heavy rain damage and typhoon damage for each local government as dependent variables. As a result, the Decision Tree model had the highest accuracy with an F1-Score of 0.56. We believe that this developed DSCM can help identify disaster risk at each stage and contribute to reducing damage through efficient disaster management for local governments based on specific events.

키워드

과제정보

본 결과물은 환경부의 재원으로 한국환경산업기술원의 물관리연구사업의 지원을 받아 연구되었습니다(21AWMP-B121100-06).

참고문헌

  1. Abrahart, R., Kneale, P.E., and See, L.M. (2004). Neural networks for hydrological modeling. CRC Press, Bock Raton, FL, U.S., pp. 1-13.
  2. Alencar, R. (2018). Resampling strategies for imbalanced datasets, accessed 9 March 2023, .
  3. AON (2021). Weather, climate & catastrophe insignt. 2018 Annual Report, London, UK.
  4. Assem, H., Ghariba, S., Makrai, G., Johnston, P., Gill, L., and Pilla, F. (2017). "Urban water flow and water level prediction based on deep learning." ECML PKDD 2017, Springer, Skopje, Macedonia, Part III, No.10, pp. 317-329.
  5. Bae, Y.H., Kim, J.S., Wang, W.J., Yoo, Y.H., Jung, J.W., and Kim, H.S. (2019). "Monthly inflow forecasting of Soyang River dam using VARMA and machine learning models." Journal of Climate Research, Vol. 14, No. 3, pp. 183-198. https://doi.org/10.14383/cri.2019.14.3.183
  6. Breiman, L. (2001). "Random forests." Machine Learning, Vol. 45, No. 1, pp. 5-32. https://doi.org/10.1023/A:1010933404324
  7. Breiman, L., and Ihaka, R. (1984). Nonlinear discriminant analysis via scaling and ACE. Department of Statistics, University of California, CA, U.S.
  8. Chawla, N.V., Bowyer, K.W., Hall, L.O., and Kegelmeyer, W.P. (2002). "SMOTE: Synthetic minority over-sampling technique." Journal of Artificial Intelligence Research, Vol. 16, pp. 321-357. https://doi.org/10.1613/jair.953
  9. Chen, T., and Guestrin, C. (2016). "Xgboost: A scalable tree boosting system." In Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, ACM, San Francisco, CA, U.S., pp. 785-794.
  10. Choi, C., Kim, J., Han, H., Han, D., and Kim, H.S. (2019). "Development of water level prediction models using machine learning in wetlands: A case study of Upo wetland in South Korea." Water, Vol. 12, No. 1, pp. 93-110. https://doi.org/10.3390/w12010093
  11. Choi, C., Kim, J., Kim, J., Kim, D., Bae, Y., and Kim, H.S. (2018a). "Development of heavy rain damage prediction model using machine learning based on big data." Advances in meteorology, Vol. 2018, 5024930.
  12. Choi, C.H. (2016). Mega flood simualtion occurred by consecutive extreme storm event and typhoon. Master Thesis, Inha University, pp. 29-31.
  13. Choi, C.H. (2019). Development of combined heavy rain damage prediction models using machine learning and effectiveness of disaster prevention projects. Ph.D. Dissertation, Inha University, pp. 1-12.
  14. Choi, C.H., Kim, J.S., Kim, D.H., Lee, J.H., Kim, D.H., and Kim, H.S. (2018b). "Development of heavy rain damage prediction functions in the seoul capital Area using machine learning techniques." Journal of The Korean Society of Hazard Mitigation, Vol. 18, No. 7, pp. 435-447. https://doi.org/10.9798/KOSHAM.2018.18.7.435
  15. Cortes, C., and Vapnik, V. (1995). "Support-vector networks." Machine Learning, Vol. 20, No. 3, pp. 273-297. https://doi.org/10.1007/BF00994018
  16. Ghumman, A.R., Ghazaw, Y.M., Sohail, A.R., and Watanabe, K. (2011). "Runoff forecasting by artificial neural network and conventional model." Alexandria Engineering Journal, Vol. 50, No. 4, pp. 345-350. https://doi.org/10.1016/j.aej.2012.01.005
  17. Go, C.M., Jeong, Y.Y., Jee, Y.G., Lee, Y.M., Kim, B.S. (2020). "A study on hydrological rainfall adjustment using machine learning and probability matching method during heavy rainfall season." Journal of Climate Research, Vol. 15, No. 4, pp. 257-267. https://doi.org/10.14383/cri.2020.15.4.257
  18. Granata, F., Gargano, R., and De Marinis, G. (2016). "Support vector regression for rainfall-runoff modeling in urban drainage: A comparison with the EPA's storm water management model." Water, Vol. 8, No. 3, 69.
  19. Han, H., Wang, W.Y., and Mao, B.H. (2005). "Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning." ICIC 2005, Springer, Hefei, China, Part 1, pp. 878-887.
  20. Han, J.W., Kwon, H.H., and Kim, T.W. (2009). "Reliability evaluation of parameter estimation methods of probability density function for estimating probability rainfalls." Journal of the Korean Society of Hazard Mitigation, Vol. 9, No. 6, pp. 143-152.
  21. Hong, J.H., Sin, T.G., Yun, U.J., Lee, T.S., and Jo, W.C. (2005). "Roadmap of NDMS Facility DB Joint Utilization System." In Proceedings of the Korean Institute of Industrial Safety Conference, KSS, pp. 179-184.
  22. Jung, J., Han, H., Kim, K., and Kim, H.S. (2021). "Machine learningbased small hydropower potential prediction under climate change." Energies, Vol. 14, No. 12, pp. 3643-3653. https://doi.org/10.3390/en14123643
  23. Kang, D.G. (2022). A decision tree for estimating mode of the response variable. Master Thesis, Korea University, pp. 6-11.
  24. Kang, T.H. (1998). Study on the development of forecasting method for rainfall, runoff and water quality in urban stream. Ph. D. Dissertation, Kyonggi University, pp. 1-21.
  25. Karatzoglou, A., Meyer, D., and Hornik, K. (2006). "Support vector machines in R." Journal of Statistical Software, Vol. 15, pp. 1-28. 
  26. Karimi, Z. (2021). Confusion matrix, research gate, accessed 23 February 2023, .
  27. Kass, G.V. (1980). "An exploratory technique for investigating large quantities of categorical data." Journal of the Royal Statistical Society: Series C (Applied Statistics), Vol. 29, No. 2, pp. 119-127. https://doi.org/10.2307/2986296
  28. Kim, B.J., Sohn, K.T., Oh, J.H., Baik, J.S., Lee, Y.H., and Baek, H.J. (2000). "Analysis of the long-term change and extreme events of daily summer rainfall over Korea." Journal of the Korean Data Analysis Society, Vol. 20, No. 1, pp. 37-44.
  29. Kim, D., Lee, J., Kim, J., Lee, M., Wang, W., and Kim, H.S. (2022a). "Comparative analysis of long short-term memory and storage function model for flood water level forecasting of Bokha stream in NamHan River, Korea." Journal of Hydrology, Vol. 606, 127415.
  30. Kim, D.H. (2018). Development of consecutive storm event based (conseb) rainfall-runoff model for short term runoff simulation and its applicability under climate change. Ph. D. Dissertation, Inha University, pp. 1-6.
  31. Kim, D.H. (2022). Development of flood water level forecasting and flood damage risk assessment method for river basin using AI-based hybrid moded. Ph. D. Dissertation, Inha University, pp. 1-173.
  32. Kim, D.H., Kim, J.W., Kwak, J.W., Necesito, I.V., Kim, J.S., and Kim, H.S. (2020). "Development of water level prediction models using deep neural network in mountain wetlands." Journal of Wetlands Research, Vol. 22, No. 2, pp. 106-112.
  33. Kim, D.H., Lee, K.S., Hwang-Bo, J.G., Kim, H.S., and Kim, S.J. (2022b). "Development of the method for flood water level forecasting and flood damage warning using an AI-based model." Journal of the Korean Society of Hazard Mitigation, Vol. 22, No. 4, pp. 145-156. https://doi.org/10.9798/KOSHAM.2022.22.4.145
  34. Kim, J.S. (2021). Development of prediction and warning technique of heavy rain damage risk based on ensemble machine learning and risk matrix. Ph. D. Dissertation, Inha University, pp. 238-242.
  35. Kim, J.S., Lee, J.H., Kim, D.H., Choi, C.H., Lee, M.J., and Kim, H.S. (2019). "Developing a prediction model (Heavy rain damage occurrence probability) based on machine learning." Journal of the Korean Society of Hazard Mitigation, Vol. 19, No. 6, pp. 115-127. https://doi.org/10.9798/KOSHAM.2019.19.6.115
  36. Kim, K.S. (2010). A study on the real time forecasting for monthly inflow Daecheong dam using hydrologic time series analyses. Master Thesis, Seokyeong University, pp.1-27.
  37. Kim, Y.H., Choi, D.Y., Jang, D.E., Yoo, H.D., and Jin, G.B. (2011). "An improvement on the criteria of special weather report for heavy rain considering the possibility of rainfall damage and the recent meteorological characteristics." Atmosphere, Vol. 21, No. 4, pp. 481-495. https://doi.org/10.14191/ATMOS.2011.21.4.481
  38. Korea Meteorological Administration (KMA) (2022). Spcial weather reports standards, accessed 27 December 2022, .
  39. Kulkarni, A., Chong, D., and Batarseh, F.A. (2020). Foundations of data imbalance and solutions for a data democracy. Academic Press, Cambridge, MA, U.S., pp. 83-106.
  40. Lee, H., Kim, H.S., Kim, S., Kim, D., and Kim, J. (2021). "Development of a method for urban flooding detection using unstructured data and deep learing." Journal of Korea Water Resources Association, Vol. 12, No. 54, pp. 1233-1242.
  41. Lee, J.S. (2021). Development and application of artificial intelligence based model for real time flood. Ph. D. Dissertation, Inha University, pp. 40-41.
  42. Liaw, A., and Wiener, M. (2002). "Classification and regression by randomForest." R News, Vol. 12, No. 3, pp. 18-22.
  43. Montanari, A., Rosso, R., and Taqqu, M.S. (1997). "Fractionally differenced ARIMA models applied to hydrologic time series: Identification, estimation, and simulation." Water Resources Research, Vol. 33, No. 5, pp. 1035-1044. https://doi.org/10.1029/97WR00043
  44. Mosavi, A., Ozturk, P., and Chau, K.W. (2018). "Flood prediction using machine learning models: Literature review." Water, Vol. 10, No. 11, 1536.
  45. Prakash, D.B., Kumar, K.A., and Kumar, R.P. (2022). "Hyper-parameter optimization using metaheuristic algorithms." CVR Journal of Science and Technology, Vol. 23, No. 1, pp. 37-43.
  46. Quinlan, J.R. (1986). "Induction of decision trees." Machine Learning, Vol. 1, pp. 81-106. https://doi.org/10.1007/BF00116251
  47. Quinlan, J.R. (1987). "Simplifying decision trees." International Journal of Man-Machine Studies, Vol. 27, No. 3, pp. 221-234. https://doi.org/10.1016/S0020-7373(87)80053-6
  48. Riad, S., Mania, J., Bouchaou, L., and Najjar, Y. (2004). "Predicting catchment flow in a semi-arid region via an artificial neural network technique." Hydrological Processes, Vol. 18, No. 13, pp. 2387-2393. https://doi.org/10.1002/hyp.1469
  49. Ryu, S.E., Shin, D.H., and Chung, K. (2020). "Prediction model of dementia risk based on XGBoost using derived variable extraction and hyper parameter optimization." IEEE Access, No. 8, pp. 177708-177720.
  50. Sharma, D.K., Chatterjee, M., Kaur, G., and Vavilala, S. (2022). Deep learning applications for disease diagnosis. Academic Press, Cambridge, MA, U.S., pp. 31-51.
  51. Shin, J.Y., Lim, S.M., Kim, J.H., and Kim, T.W. (2014). "Analysis of urban flood damage characteristics using inland flood scenarios and flood damage curve." Journal of the Korean Society of Hazard Mitigation, Vol. 14, No. 1, pp. 291-302. https://doi.org/10.9798/KOSHAM.2014.14.1.291
  52. Shoaib, M., Shamseldin, A.Y., Melville, B.W., Khan, M.M. (2016). "A comparison between wavelet based static and dynamic neural network approaches for runoff prediction." Journal of Hydrology, Vol. 535, pp. 211-225. https://doi.org/10.1016/j.jhydrol.2016.01.076
  53. Song, Y.S., and Chae, B.G. (2008). "Development to prediction technique of slope hazards in gneiss area using decision tree model." The Journal of Engineering Geology, Vol. 18, No. 1, pp. 45-54.
  54. Yan, J., Jin, J., Chen, F., Yu, G., Yin, H., and Wang, W. (2018). "Urban flash flood forecast using support vector machine and numerical simulation." Journal of Hydroinformatics, Vol. 21, No. 1, 016111.