Abstract:By combining chaotic mapping and adaptive inertial weight, we improve the standard whale algorithm so as to improve the global optimization ability and convergence speed of the algorithm. Aiming at the disadvantage of BP neural network, we use the improved whale algorithm to optimize BP neural network. On this basis, we establish the improved whale algorithm to optimize the BP neural network GPS elevation anomaly fitting prediction model, and the model is verified by two groups of GPS data in different terrain feature engineering. The results show that the BP model optimized by the improved whale algorithm can achieve higher accuracy and stability in GPS height fitting.