基于AHP-物元可拓的重庆市农房安全评估模型研究OA

Rural housing safety is a critical concern in disaster-prone mountainous regions,such as Chongqing,China.Traditional assessment methods are often inefficient and subjective,struggling to integrate multi-source risk factors including structural conditions,environmental hazards,and human management practic-es.In this study,a quantitative,objective,and practical safety assessment model is developed for rural houses by integrating the Analytic Hierarchy Process(AHP)and matter-element extension theory,aiming to facilitate rapid screening and tiered management for disaster mitigation.A three-dimensional assessment framework is estab-lished,comprising Structural Stability(7 indicators),Environmental Risk Exposure(5 indicators),and Manage-ment Compliance(3 indicators),totaling 15 indicators.indicator weights are determined via a three-step hybrid approach:①objective pre-ranking using the simple correlation function of matter-element extension theory;②subjective pairwise comparisons by 20 experts to construct AHP judgment matrices;and ③dynamic weight modi-fication using expert validity coefficients derived from the matter-element distance between individual judgments and group consensus.A fuzzy quantification standard is developed to transform all indicator values into a compa-rable[0,1]scale,with higher values indicating a safer state.The comprehensive safety index is then calculated through weighted aggregation and classified into four levels:Safe(N≥0.6),Basically Safe(0.5≤N<0.6),At-risk(0.4≤N<0.5),and Dangerous(N<0.4).The model is empirically validated on 1,290 rural houses sampled from the Dazu,Banan,and Wanzhou districts of Chongqing,Results are benchmarked against on-site profession-al assessments following the Chongqing Technical Guidelines(YU Jian Cun Zhen[2020]No.5).The model demonstrated high consistency with on-site judgments.The overall agreement rate is 95.2%,with a Cohen's Kappa coefficient of 0.853(p<0.001),indicating almost perfect agreement.For the Safe level,the model achieved a precision of 98.2%and a recall of 98.3%(F1-score=0.983),proving its efficacy in rapid screen-ing.Crucially,for the Dangerous level,the model achieved 100%recall,identifying all high-risk houses without omission,although precision is moderate at 57.1%due to some At-risk houses being misclassified as Dangerous.From a disaster prevention perspective,100%recall is paramount as the cost of omission far exceeds false alarms.For the intermediate levels(Basically Safe and At-risk),recalls are 73.9%and 87.5%respectively,re-flecting the inherent difficulty in classifying boundary samples.Notably,all 48 houses classified by the model as At-risk or Dangerous are confirmed as B,C,or D grade on-site,with none being A-grade(Safe),demonstrating the model's effectiveness in identifying houses requiring attention.The proposed AHP-matter-element extension model provides a robust,objective,and standardized tool for rural house safety assessment in Chongqing's mountainous areas.Its key innovations lie in the hybrid weight determination method integrating objective corre-lation and expert validity,and the comprehensive indicator system covering structural,environmental,and mana-gerial factors.The weight distribution shows Structural Stability(0.595 5)as the most critical dimension,fol-lowed by Environmental Risk Exposure(0.296 3)and Management Compliance(0.108 2),reflecting that struc-tural integrity forms the foundation of safety while environmental threats and human interventions play crucial modulating roles.The model effectively identifies safe houses for exemption from detailed inspection and flags high-risk houses for priority intervention.Future work should focus on refining quantification rules with empiri-cal data,exploring adaptive weight adjustment mechanisms,and validating the model's generalizability in other geographical contexts.