Abstract：In view of the non-linearity and non-stationary characteristics of satellite clock bias(SCB) time series, as well as the interference between trend and noise components that may affect the accuracy of prediction, this paper proposes a SCB prediction model(SSA-ANFIS) based on singular spectrum analysis(SSA) and adaptive neuro-fuzzy inference system(ANFIS). This paper first uses SSA to decompose and reconstruct the first-order difference sequence of clock bias, obtaining the trend component and the residual component. Then, it uses the ANFIS model to predict the reconstructed components, and superimposes and restores the predicted results to obtain the final predicted clock bias value. Finally, through experiments, this paper compares the proposed model with GM, QP, LSTM and ANFIS models. The results show that SSA-ANFIS model can effectively improve the prediction accuracy of the single model. Compared with the LSTM and ANFIS models, its prediction accuracy increased by 25.7%-40.7% and 39.4%-45.7%, respectively.
