基于Logistic回归的社会单位消防安全管理水平分析

Analysis of fire safety management level of social units based on logistic regression

针对不同高度建筑及重点监管场所需实施不同等级防火监管措施的问题,提出基于无序多分类Logistic回归模型的防火安全管理水平分析策略。对低层/多层建筑、高层建筑和超高层建筑分别进行定性分析,然后通过R语言建立无序多分类Logistic回归模型进行定量分析。考虑到人员密集场所有必要给予重点分析,归纳总结了各类人员密集场所并对其建立回归模型进行防火监管水平分析。最后,对所有建筑样本的综合管理水平进行Logistic回归分析,进而给出有针对性的管理措施和建议。

For the problem of different heights of buildings and key supervisory places need to implement different levels of fire prevention supervision measures,the fire safety management level analysis strategy based on disordered multiclassification logistic regression model is proposed.Qualitative analysis is conducted for low-rise/multi-story buildings,high-rise buildings and super high-rise buildings respectively,and then a disordered multi-classification logistic regression model is established for quantitative analysis through R language.Considering that it is necessary to give focus to the analysis of crowded places,various types of crowded places are summarized and regression models are established for their fire prevention supervision level analysis.Finally,the comprehensive management level of all building samples is analyzed by logistic regression,and then targeted management measures and suggestions are given.