Predicting the High-rise building construction project safety risk based on QPSO-SVM model

In this paper, we aim to solve the problem of high-rise building construction project safety risk forecasting. The main innovation of this paper lies in that we convert the project safety risk forecasting problem to a classification problem, and design a novel QPSO-SVM model to implement the classification process. Before predicting the project safety risk, we present an index system, which contains six first-layer indexes, such as “Falling accident”, “Objects striking accident”, “Collapse accident”, “Mechanical injury”, “Fire disaster in construction”, and “Electric shock injury”. Particularly, 29 indexes are included in the second-layer of this index system, and these 29 indexes can effective represent almost all influencing factors in high-rise building construction project safety risk prediction. Next, we describe how to select optimal parameters of SVM classifier using the quantum behaved particle swarm optimization policy. Finally, experimental results demonstrate that, compared with other schemes, the proposed hybrid QPSO-SVM can forecast the high-rise building construction project safety risk with higher accuracy.