Prediction of foundation pit surface settlement based on SSA-ELM algorithm
The traditional extreme learning machine(ELM)algorithm is prone to fall into local minimum,inac-curate parameter selection in network structure and poor prediction accuracy when it is used to predict the sur-face settlement of deep foundation pit.A prediction model of foundation pit surface settlement based on sparrow search algorithm(SSA)optimized extreme learning machine algorithm is proposed.According to the characteris-tics of sparrow search algorithm,such as fast convergence speed,strong optimization ability and stability,the connection weights and thresholds in the extreme learning machine algorithm are optimized.The optimized mod-el is applied to the prediction of foundation pit surface settlement.The prediction accuracy of the sparrow search algorithm optimized extreme learning machine algorithm(SSA-ELM)is compared with the traditional ELM algo-rithm,GA-ELM and PSO-ELM algorithms.The experimental results show that the prediction accuracy of SSA-ELM algorithm is higher than the traditional ELM algorithm,GA-ELM algorithm and PSO-ELM algorithm,and the SSA-ELM algorithm is more stable,which has better effect in the prediction of ground settlement of the foun-dation pit.The SSA-ELM algorithm has achieved the purpose of improving the accuracy of prediction,and has certain feasibility and practicality.
万方数据
维普
