Modeling of Direct Economic Losses of Storm Surge Disasters Based on a Novel Hybrid Forecasting System

Abstract

Modeling the direct economic losses of storm surge disasters can assess the disaster situation in a timely manner and improve the efficiency of post-disaster management in practice, which is acknowledged as one of the most significant issues in clean production. However, improving the forecasting accuracy of direct economic losses caused by storm surge disasters remains challenging, which is also a major concern in the field of disaster risk management. In particular, most of the previous studies have mainly focused on individual models, which ignored the significance of reduction and optimization. Therefore, a novel direct economic loss forecasting system for storm surge disasters is proposed in this study, which includes reduction, forecasting, and evaluation modules. In this system, a forecasting module based on an improved machine learning technique is proposed, which improves the generalization ability and robustness of the system. In addition, the key attributes and samples are selected by the proposed reduction module to further improve the forecasting performance from the two innovative perspectives. Moreover, an evaluation module is incorporated to comprehensively evaluate the superiority of the developed forecasting system. Data on the storm surge disasters from three typical provinces are utilized to conduct a case study, and the performance of the proposed forecasting system is analyzed and compared with eight comparison models. The experimental results show that the mean absolute percentage error (MAPE) predicted by the Extreme Learning Machine (ELM) model was 16.5293%, and the MAPE predicted by the proposed system was 1.0313%. Overall, the results show that the performance of the proposed forecasting system is superior compared to other models, and it is suitable for the forecasting of direct economic losses resulting from storm surge disasters.