基于BAS-BP模型的风暴潮灾害损失预测

风暴潮灾害是中国沿海地区最严重的灾害之一，近年来由其带来的经济损失均占海洋灾害总损失的90%以上，因此构建一个简单准确的损失预估模型显得尤为重要。本文以现有风暴潮灾害研究为基础建立了基于天牛须搜索（beetle antennae search）优化的BP神经网络模型，将其应用到风暴潮灾害经济损失预评估中。本文收集了福建省1994~2016年记录比较完善的29个风暴潮灾害损失数据，建立风暴潮灾害损失预评估指标体系并利用熵值法对指标因子进行预处理，消除数据冗余信息对预测的影响。对模型进行仿真测试，结果表明，与标准BP神经网络相比新模型有效避免了网络陷入局部极小值的可能，且与常规优化算法相比，克服了训练时间长、收敛速度慢的缺点，具有更好的鲁棒性和预测精度。

The assessment of storm surge disaster loss based on BAS-BP model

Storm surge is one of the most devastating coastal disasters in China's coastal areas.In recent years economic losses of storm surge accounted for more than 90 percent of the total loss of marine disasters.Therefore, it is very important to develop a simple and accurate projection and assessment model.On the basis of the existing storm surge disaster the BP neural network model based on the optimization of beetle antennae search algorithm is established, and apply it to the prediction of economic damage of storm surge.This paper collected data of 29 storm surge disasters in Fujian province from 1994 to 2016.The evaluation index system of storm surge disaster is established and the index factors are preprocessed by entropy method to eliminate redundancy information.The simulation results show that compared with the standard BP neural network, the new model effectively avoids the possibility of network getting into local minimum, and compared with the conventional optimization algorithm, the disadvantages of long training time and slow convergence speed are overcome.The BAS-BP model performs more robust and accurate.