论文部分内容阅读
讨论了利用BP神经网络,对震后的死亡人数进行较为准确的评估,进而指导后续的地震应急工作。选取地震震级、震源深度、震中烈度、抗震设防烈度、震中烈度与抗震设防烈度之差、发震地人口密度和时间(白天、晚上或凌晨)等7项数据作为输入因子,以死亡人数作为输出因子,建立了BP神经网络模型,再以阳江、炉霍、鲁甸和景谷4次地震的该7项数据作为检测因子检测建立的神经网络。可以得出:由网络训练得到的结果与真实的死亡人数相差很小,结果对比比较理想。
Discusses the use of BP neural network to make a more accurate assessment of the post-earthquake death toll, and then guide the follow-up earthquake emergency work. Seismic intensity, focal depth, epicenter intensity, seismic fortification intensity, difference between epicenter intensity and seismic fortification intensity, population density and time of seismogenic area (daytime, evening or early morning) are selected as input factors and the death rate is taken as output Factor, a BP neural network model is established, and then the neural network established by using the seven data of Yangjiang, Luhuo, Ludian and Jinggu 4 earthquakes as detection factors. It can be concluded that the result obtained by the network training is quite different from the real death number, and the result is ideal.