基于SSA-SVM的道路事故预测模型

Road Accident Prediction Model Based on SSA-SVM

将道路事故特征因素和环境因素作为自变量,事故严重程度作为因变量,采用SVM构建道路事故预测模型,并使用SSA优化模型中的最优参数组合,最终建立基于SSA优化的SVM道路事故预测模型,模型预测准确率优于广义线性预测模型和BP神经网络预测模型。经过模型致因分析,得到三项对道路事故严重程度具有显著影响的自变量,分别为机非隔离形式、监控设施和最近交叉口交通组织形式,可作为道路交通安全管控的核心控制参数。

With road accident characteristic factors and environmental factors as independent variables and accident severity as dependent variables,SVM was used to build a road accident prediction model,and the optimal parameter combination in the SSA optimization model was used to establish a road accident prediction model based on SSA optimization.The prediction accuracy of the model was better than that of generalized linear prediction model and BP neural network prediction model.Through the model causation analysis,three independent variables that have a significant impact on the severity of road accidents are obtained,namely,the non-isolation form,monitoring facilities and the traffic organization form of the nearest intersection,which can be used as the core control parameters of road traffic safety management and control.