Analysis of Typhoon Path for the Korean Peninsula using Machine Learning Algorithms
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This research enhances typhoon path prediction by integrating advanced machine learning with extensive meteorological data, focusing on the Korean Peninsula. Utilizing a novel two-step modeling approach, it predicts if typhoons will cross latitude 33 degrees and their paths thereafter, categorizing them into four types based on their trajectories relative to the Peninsula. Models were assessed using data from 1979 to 2023, employing variable importance and partial dependence plots to decipher the decision-making processes. Findings show that longitude and meteorological factors like air and skin temperature significantly affect typhoon paths. The study sets a new standard in typhoon forecasting by substituting high-dimensional spatio-temporal data with localized datasets and offers insights for disaster management, setting a new standard in typhoon forecasting.
Abstract
This research enhances typhoon path prediction by integrating advanced machine learning with extensive meteorological data, focusing on the Korean Peninsula. Utilizing a novel two-step modeling approach, it predicts if typhoons will cross latitude 33 degrees and their paths thereafter, categorizing them into four types based on their trajectories relative to the Peninsula. Models were assessed using data from 1979 to 2023, employing variable importance and partial dependence plots to decipher the decision-making processes. Findings show that longitude and meteorological factors like air and skin temperature significantly affect typhoon paths. The study sets a new standard in typhoon forecasting by substituting high-dimensional spatio-temporal data with localized datasets and offers insights for disaster management, setting a new standard in typhoon forecasting.
참고문헌 (14)
- Ahn, Suk-Hee, "The characteristics of disaster by track of typhoon affecting the Korean Peninsula", Journal of the Korean Society of Hazard Mitigation, 8(3), pp.29~36 (2008)
- Breiman, Leo, "Bagging predictors", Machine Learning, 24, pp.123~140 (1996)
- Breiman, Leo, "Random forests", Machine Learning, 45, pp.5~32 (2001)
- Chen, Rui, "A hybrid CNN-LSTM model for typhoon formation forecasting", GeoInformatica, 23, pp.375~396 (2019)
- Chen, Tianqi, "Xgboost: A scalable tree boosting system", Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data mining (2016)
- Friedman, Jerome H., "Greedy function approximation: a gradient boosting machine", Annals of Statistics, pp.1189~1232 (2001)
- Jiang, Shuailong, "Improvement of typhoon intensity forecasting by using a novel spatio-temporal deep learning model", Remote Sensing, 14(20), pp.5205~ (2022)
- Kalnay, E., "The NMC", NCAR 40-year reanalysis project, 1996, pp.77(3), 437~472
- Kanamitsu, Masao, "Ncep–doe amip-ii reanalysis(r-2)", Bulletin of the American Meteorological Society, 83(11), pp.1631~1644 (2002)
- Ke, Guolin, "Lightgbm : A highly efficient gradient boosting decision tree", Advances in Neural Information Processing Systems, 30 (2017)
- Kursa, Miron B., "Boruta–a system for feature selection", Fundamenta Informaticae, 101(4), pp.271~285 (2010)
- Prokhorenkova, Liudmila, "CatBoost : unbiased boosting with categorical features", Advances in Neural Information Processing Systems, 31 (2018)
- Quinlan, J. Ross, "Induction of decision trees", Machine Learning, 1, pp.81~106 (1986)
- Xu, Guangning, "AM-ConvGRU : A spatio-temporal model for typhoon path prediction", Neural Computing and Applications, 34(8), pp.5905~5921 (2022)