摘要
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、及时施救和防灾预案编制有举足轻重的指导意义。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,采用了S 函数输出限幅,引入了惯性因子,加快了学习的收敛速度,提高了预测的精度。
City flood forecast is the important content of loss reduction for insurance agent. It can forecast the stricken area and its intension, which has significance in making rate, preventing efficiently, rescuing in time and working out the prediction. Adopting BP algorithm, combining clustering and importing inertia gene accelerate the speed of learning and improve the precision of forecast.
出处
《计算机工程与设计》
CSCD
北大核心
2005年第3期699-701,734,共4页
Computer Engineering and Design