| 三层BP神经网络地震灾害人员伤亡预测模型 |
| |
| 引用本文： | 于山,王海霞,马亚杰.三层BP神经网络地震灾害人员伤亡预测模型[J].地震工程与工程振动,2005,25(6):113-117. |
| |
| 作者姓名： | 于山 王海霞 马亚杰 |
| |
| 作者单位： | 天津大学,管理学院,天津,300073;河北理工大学,建工学院,河北省地震工程研究中心,河北,唐山,063009 |
| |
| 摘 要： | 选择地震发生时刻、震级、震中烈度、建筑物倒塌和严重破坏率、抗震设防水准、人口密度、地震预报等7个评价指标，以20次严重地震灾害为示例（其中，17个作训练样本，3个作验证样本），建立了三层BP神经网络地震灾害人员伤亡预测模型。基于MATLAB6，5BP神经网络训练，得出的预测结果与各个示例的实际数值比较吻合。验证样本的训练结果表明，该模型适用于地震灾害人员伤亡评估。通过对评价指标的权重计算，确认人口密度、建筑物倒塌与严重破坏率、震中烈度是影响地震灾害人员伤亡的主要因素，地震预报、抗震设防水准、地震发生时刻和震级次之。作为人为可控预测指标，减少人口密度特别是城市人口密度，提高建（构）筑物抗震能力及预测预报水平，对于减少地震灾害人员伤亡起更重要的作用。
|
| 关 键 词： | 神经网络 地震灾害 人员伤亡 预测模型 |
| 文章编号： | 1000-1301（2005）06-0113-05 |
| 收稿时间： | 2005-06-21 |
| 修稿时间： | 2005-06-212005-10-19 |
Three-layer BP network model for estimation of casualties in an earthquake |
| |
| Yu Shan,Wang Hai-xia,Ma Ya-jie.Three-layer BP network model for estimation of casualties in an earthquake[J].Earthquake Engineering and Engineering Vibration,2005,25(6):113-117. |
| |
| Authors: | Yu Shan Wang Hai-xia Ma Ya-jie |
| |
| Institution: | 1. School of Management, Tianjin University, Tianjin 300073 ,China; 2. School of Architecture and Civil Engineering, Hebie Polytechnic University, Earthquake Engineering, Research Center of Hebei Province, Tangshan 063009,China |
| |
| Abstract: | In this paper,a three-layer BP network model for estimating the casualties in earthquakes is presented,in which seven evaluating indices of earthquake such as time of occurrence,magnitude,epicentral intensity,collapse of buildings,level of protection,population density and disaster forecasting are included and the information of 20 past earthquake disasters is quoted as samples,in which 17 samples are used as training model and 3 samples as verification model.With the use of ANN tool in Matlab6.5P,the predicting result of the training model well matches the practical data of verifying samples.And the results of verifying samples demonstrate the adaptability of the model for estimating the casualties in earthquakes.Considering the weights of the indices in the model,it is revealed that the population density,the destructiveness of buildings and the epicentral intensity of earthquake are the main factors that influence the casualties in earthquakes,while the rest of the indices are secondary factors.In these factors,reducing population density(especially city population density),improving earthquake-proof capability of buildings and enhancing forecasting accuracy,which could be affected by human activities,are more important for reducing the casualties in earthquakes. |
| |
| Keywords: | ANN earthquake casualties predictive model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
