Gray relational analysis and SBOA-BP for predicting settlement intervals of high-speed railway subgrade

Purpose The deformation of the roadbed is easily influenced by the external environment to improve the accuracy of high-speed railway subgrade settlement prediction. Design/methodology/approach A high-speed railway subgrade settlement interval prediction method using the secretary bird optimization (SBOA) algorithm to optimize the BP neural network under the premise of gray relational analysis is proposed. Findings Using the SBOA algorithm to optimize the BP neural network, the optimal weights and thresholds are obtained, and the best parameter prediction model is combined. The data were collected from the sensors deployed through the subgrade settlement monitoring system, and the gray relational analysis is used to verify that all four influencing factors had a great correlation to the subgrade settlement, and the collected data are verified using the model. Originality/value The experimental results show that the SBOA-BP model has higher prediction accuracy than the BP model, and the SBOA-BP model has a wider range of prediction intervals for a given confidence level, which can provide higher guiding value for practical engineering applications.