EOP prediction using least square fitting and autoregressive filter over optimized data intervals

This study firstly employs the calculation of base sequence with different length, in 1–90 day predictions of EOP (the UT1-UTC and polar motion), by the combined method of least squares and autoregressive model, and find the base sequence with best result for different prediction spans, which we call as “predictions over optimized data intervals”. Compared to the EOP predictions with fixed base data intervals, the “predictions over optimized data intervals” performs better for the prediction of UT1-UTC, and shows a significant improvement for the prediction of polar motion, and particularly promotes our competitive level in the international activity of Earth Orientation Parameters Combination of Prediction Pilot Project.