Developing a Prediction Model (Heavy Rain Damage Occurrence Probability) Based on Machine Learning
Authors:
- Jongsung KimJKJongsung Kim
- Donghyun KimDKDonghyun Kim
- Changhyun ChoiCCChanghyun Choi
- Myungjin LeeMLMyungjin Lee
- Hung Soo KimHung Soo Kim
Open Access
- 30 November 2019
- journal article
- Published by Korean Society of Hazard Mitigation in Journal of the Korean Society of Hazard Mitigation
- Vol. 19 (6) , 115-127
Abstract
In this study, a prediction model (heavy rain damage occurrence probability or PM-HDOP) was developed for a metropolitan area. The heavy rain damage and rainfall data were collected as dependent and independent variables, respectively. The dataset was divided into training (2005-2016) and test sections (2017). We developed the PM-HDOP using machine learning methods such as logistic regression, artificial neural network, bagging, random forest, and boosting to predict the occurrence of nonlinear natural disasters. An architectural model with the best performance was selected, and the PM-HDOP was subsequently used to predict the probability of occurrence. As a result, a boosting scheme showed the best performance in Gyeonggi-do and Seoul, and a bagging scheme showed the best performance in Incheon. If the results of this study are used to predict the occurrence of heavy rain damage, which is not currently being serviced in Korea, it is possible to effectively reduce the damages.Keywords
Funding Information
- Ministry of the Interior and Safety (재난-2015-05)
This publication has 24 references indexed in Scilit:
- Development of Heavy Rain Damage Prediction Model Using Machine Learning Based on Big DataAdvances in Meteorology, 2018
- Rainfall prediction using hybrid neural network approach2018
- An extensive evaluation of seven machine learning methods for rainfall prediction in weather derivativesExpert Systems with Applications, 2017
- Development of Heavy Rain Damage Prediction Function Using Statistical MethodologyJournal of the Korean Society of Hazard Mitigation, 2017
- A rainfall prediction model using artificial neural network2012
- Application of artificial neural networks to rainfall forecasting in Queensland, AustraliaAdvances in Atmospheric Sciences, 2012
- Landslide characteristics and slope instability modeling using GIS, Lantau Island, Hong KongGeomorphology, 2002
- Vulnerability of the Netherlands and Northwest Europe to Storm Damage under Climate ChangeClimatic Change, 1999
- Artificial neural networks, back propagation, and the Kelley-Bryson gradient procedureJournal of Guidance, Control, and Dynamics, 1990
- The Regression Analysis of Binary SequencesJournal of the Royal Statistical Society Series B: Statistical Methodology, 1958