Construction pit deformation measurement technology based on neural network algorithm

Abstract The current technology of foundation pit deformation measurement is inefficient, and its accuracy not ideal. Therefore, an intelligent prediction model based on back propagation neural network (BPNN) proposed to predict the intelligently, with high efficiency, so as improve safety project. Firstly, address shortcomings BPNNs, which rely initial parameter settings tend fall into local optimum unstable performance, this study adopts modified particle swarm optimization (MPSO) optimise parameters BPNNs constructs a MPSO–BP algorithm achieve predictive measurements deformation. After training testing data samples, results show that 99.76%, 2.25% higher than optimization–back (PSO–BP) 3.01% BP model. aforementioned in can effectively variables construction projects provide support for protective measures staff, helpful cause China.