Forecasting direct economic losses of marine disasters in China based on a novel combined model. (December 2020)

• Record Type: Journal Article
• Authors: Zhao, Xin; Li, Hui; Ding, Lili; Wang, Wei; Xue, Yuemei
• Is Part Of: International journal of disaster risk reduction. Volume 51(2020)
• Publication Date: 2020-12
• Subjects: Direct economic losses of marine disasters -- Definite integral -- Marine disaster index -- Neural network; Emergency management -- Periodicals; Risk management -- Periodicals; Disaster relief -- Periodicals; Hazard mitigation -- Periodicals; 363.34
• DOI: 10.1016/j.ijdrr.2020.101921
• Languages: English
• ISSNs: 2212-4209

Abstract

Abstract: The accurate prediction of marine disaster losses is necessary to improve the efficiency of policy makers and researchers. This paper proposes a novel model, i.e., the ENN-GRNN-DI model, combining the Elman neural network with a generalized regression neural network and definite integral to forecast the direct economic losses of marine disasters. The proposed combined model not only makes full use of the data information, but it also achieves interval prediction that obviously increases the reliability and enhances the forecasting performance. To evaluate the performance, a comparison of the proposed combined model and the conventional models is provided. The experimental results show that the novel combined model outperforms the benchmark models in forecasting the direct economic losses of marine disasters. Highlights: TThis study takes advantage of the characteristics of DELMD and applies ENN in weak forecasting. This study employs the GRNN to boost the ENN and avoid the effects of overfitting. The definite integral is employed to obtain an interval prediction, which is whmore instructive in marine disaster management. The experimental results show that the novel combined model outperforms the benchmark models in forecasting the DELMD.