基于多因素的建筑安全GA-BP神经网络事故预测模型研究

Research on Prediction Model of GA-BP Neural Network for Building Safety Accidents based on Multiple Factors

传统的建筑安全事故预测模型一般是在历史事故的相关数据上建立的预测模型,但是它们普遍没有分析事故的成因。而本文建立的结合事故成因分析的BP神经网络预测模型是先分析近十来年的建筑安全事故成因,并采用专家评分法对事故成因指标进行量化处理,将量化处理后的因素指标作为网络的输入,将事故的死亡人数作为网络的输出,建立基于多因素的BP神经网络事故预测模型。结果发现基于多因素的BP神经网络事故预测模型比未结合事故成因分析的BP神经网络模型预测的平均相对误差降低了1.75%。为了进一步提高预测模型的精度,采用遗传算法对基于多因素的BP神经网络预测模型进行优化,优化后的预测模型平均相对误差降低了3.49%。

Traditional accident prediction models for building safety are generally based on the historical accident data,which did not analyze the causes of the accidents.The BP neural network prediction model based on accident cause analysis in this paper firstly analyzes the causes of building safety accidents in the past ten years.The causes of the accidents are quantified by the expert scoring method,which are taken as the input of the network,while the number of deaths from accidents as the output of the network.Finally,BP neural network accident prediction model based on multiple factors is established.The results show that the BP neural network accident prediction model based on multiple factors has better prediction effect,and the average relative error of prediction is reduced by 1.75%.In order to further improve the accuracy of the prediction model,genetic algorithm is used to optimize the initial weight and threshold of BP neural network,and the average relative error of the optimized prediction model was reduced by 3.49%.