基于BP神经网络的灾情预测模型

Disaster Forecast Model Based on BP Network

洪灾灾情预测是保险公司财产保险防灾减损工作的重要内容,它有效地预测出受灾的地区和强度,对于财产保险的费率制定、有效预防、及时施救以及防灾预案的编制有重要的指导意义。该文采用了BP神经网络进行灾情预测,在学习过程中结合了聚类,引入了惯性因子,加快了学习时的收敛速度,试验证明,取得了良好的学习和测试效果。

Disaster forecast is the important content of loss reduction for insurance agent. It can forecast the stricken area and its intension, which has signality in making rate, preventing efficiently, rescue in time and the workout of the prediction. This paper adopts BP arithmetic combining clustering, importing inertia gene, accelerates the speed of learning and gets the nicer result.