【摘要】 强烈的地震给人们生命财产带来巨大损失,为了能够在地震之前预测出建筑物震害,提出一多层砖房为例。利用MATLAB神经网络工具箱,建立一种基于贝叶斯正则算法的BP神经网络模型,并以过去发生地震地区的多层砖房调查数据为震害因子的震害预测方法。结果表明:对多层砖房的震害样本的预测达到理想效果。更多还原

【Abstract】 A strong earthquake causes huge losses for people’s life and property. In order to predict the building earthquake damage before earthquake,we take the multistory masonry building as the sample to propose by MATLAB neural network toolbox,we should establish one kind of BP neural network model based on Bayesian method and take the survey data of multistory masonry building as the earthquake damage prediction method of earthquake damage factor. The result indicates that the method can achieve an ideal effect for predicting earthquake damage of the multistory masonry building .更多还原