基于情报信息的空防安全威胁评估回归模型研究  

Research on the Regression Model of Civil Aviation Security Threats Based on Intelligence Information

作　　者：庄刘[1] 魏中许[2] 

机构地区：[1]中国民航飞行学院计算机学院,四川广汉618307 [2]中国民航飞行学院空防安全研究室,四川广汉618307

出　　处：《四川师范大学学报（自然科学版）》2015年第3期460-465,共6页Journal of Sichuan Normal University（Natural Science）

基　　金：国家自然科学基金(61079022)资助项目

摘　　要：空防安全威胁传统评估方法主要是基于情报信息形成专家或领导意见的定性分析方法,既缺乏威胁评估理论模型,又缺乏有效性的实证分析.首先依据定性分析考量空防安全事件威胁的维度,通过对情报信息要素进行分解,构建了空防安全威胁定量评估的指标体系;其次,以劫炸机案例模拟威胁情报信息,应用多元线性回归的方法,采用向后删除法对变量进行优选,建立了基于威胁主体属性、犯罪动机、威胁手段、机场、敏感时期五项情报信息指标的威胁评估回归模型.该模型的拟合检验结果表明,基于情报信息的威胁评估与空防安全专家评估意见具有很好的一致性.The traditional assessment approach to the aviation security threat is qualitative and mainly relies on experts＇ or leaders＇ judgments from intelligence information with a lack of theoretical guidance of threats assessment and empirical analysis.With qualitative methods,dimensions of aviation security threat and further decomposed related intelligence information were achieved to develop indicator system to quantify civil aviation security threats.Taking hijack and airplane bombing as threats to extract indicator data of intelligence information,the threat security threats assessment model including five indicators,target airport,attack means,attacker＇ s motive,subject of threat,and sensitive period or not,was developed with backward elimination regression method.The fitting test results show that the model based on intelligence information is in good consistentency with judgments of aviation security experts.

关 键 词：情报信息 空防安全威胁 多元线性回归 向后删除变量法 

分 类 号：O212.4[理学—概率论与数理统计]