大型公共活动拥挤踩踏事故诱因错综复杂,为预先获取其危险程度,本文借鉴前人研究成果,从人、物、环、管理四个方面,建立明确指标风险分级标准的评估指标体系,进而在此基础上引入遗传算法(GA)优化的支持向量机(SVM)方法,构建基于GA-SVM的大型公共活动拥挤踩踏事故风险评估模型。应用建成的风险评估模型,对实例进行评估,结果表明,该模型评估结果精确度较高,在风险评估中有较好的适用性和可靠性。 In order to gain prestige degree of dangerousness, the article draws on the research results of the predecessors and establishes the evaluation index system which defines the risk classification standard of indicators from four aspects: people, material, environment and management, and then based on this The Genetic Algorithm (GA) -sized Support Vector Machine (SVM) method is introduced to construct a GA-SVM-based model of crowded stampede risk assessment. The established risk assessment model is applied to evaluate the examples. The results show that the model has high accuracy and good applicability and reliability in risk assessment.