【摘要】 应用云南省最近几次地震中多层砖房的震害实例，构造了多层砖房震害预测的人工神经网络 方法，并利用这些数据进行了网络训练。结果表明，这种方法对已训练数据有很好的适应性，但如 果要将之应用于震害预测，则尚需积累大量的震害数据，并进行大量的网络试验工作。更多还原

【Abstract】 Based on several damage examples from the recent earthquakes occurred in Yunnan Province, an artificial neural network method is constructed for damage forecast of multistory masonry buildings. For the part of network training, it demonstrates great adaptability for these real damage data. While for the usage of network, it is not good enough for practical exercises. A lot of damage data and a number of try and error works are needed to improve this network.更多还原