CONSTRUCTION SAFETY RISK ASSESSMENT BASED ON THE PROJECTION PURSUIT MODEL

Scientific and effective evaluation of construction safety risks reduces the frequency of safety accidents. A novel construction safety evaluation model was developed based on the projection pursuit model (PP) improved by particle swarm optimization (PSO). A risk evaluation system was constructed using a literature review and expert interviews. The PP was used for data mining to obtain objective indicator weights and determine the risk level, and the PSO was used to derive the optimal projection vector of the PP. A case study of the Mu-nicipal Administration Center Project at the Nanchang Airport Economic Zone was conducted. The proposed model had higher accuracy and better resolution and provided a more objective evaluation of the construction safety risks than classical methods, such as fuzzy comprehensive evaluation and the technique for order of preference by similarity to an ideal solution (TOPSIS).