Abstract: Landslide deformation is a key evaluation parameter for judging whether the landslide is stable after treatment. Carrying out the deformation prediction of landslide can grasp the stability of the landslide, which is beneficial to the risk analysis of landslide and facilitates the prevention and control of geological disasters. To accurately predict the deformation of the treated landslide, this paper proposes a deformation predicting method of the treated landslide by using the Bird Swarm Algorithm (BSA) to optimize the BP neural network. The deformation prediction model of highway slope in Guangxi Province is established with BSA-BP neural network, and the prediction results of BSA-BP neural network and normal BP neural network are compared and analyzed. It is indicated that the mean square error and correlation coefficient of the prediction results of the BSA-BP neural network are 0.053 4 and 0.997 6, respectively. The mean square error and correlation coefficient to prediction results are 2.225 6 and 0.968 with BP neural network. Bird Swarm Algorithm can effectively improve the prediction accuracy of the BP neural network model. The BSA-BP neural network model can be effectively applied to the deformation prediction of the treated landslide. The research results could provide a reference for risk prediction of the treated landslide in the future.

Key words: landslide, BSA-BP neural network, bird swarm algorithm, deformation prediction