Prediction of Surface Settlement in Shield-Tunneling Construction Process Using PCA-PSO-RVM Machine Learning

Surface settlement is one of the key engineering issues during shield construction process. In order to accurately predict surface settlement, this paper proposes a new machine learning method based on relevance vector (RVM), principal component analysis (PCA), and particle swarm optimization (PSO). Taking Beijing Metro Line 6 as case study, PCA-PSO-RVM model used make prediction compared with results RVM using same samples. evaluate reliability model, three evaluation indexes including mean relative error (MRE), root square (RMSE), Theil inequality coefficient (TIC) were calculated, sensitivity was carried out them. The show that minimum between actual value only 0.06%. calculated MRE, RMSE, TIC are 0.17%, 0.0714 mm, 0.027%, respectively, which shows has higher accuracy, smaller deviations, other models. Through analysis, it found weighted average internal friction angle (φ) most significant impact should be focused in relevant research.