关键词: 二次多项式模型, 灰色模型, 原子钟, 数据丢失, 残差, 组合模型

Abstract: High accuracy of time scale is established and maintained by atomic clocks. During the continuous measurement, the measurement data may get lost due to the delay of the equipment or the sudden change in the outside environment. The quadratic polynomial model can be used to predict and correct the lost data, but it’s not good when dealing with continuous lost data. This paper uses grey model and another two combined models to predict and correct the lost data, and also compares the four models’ prediction accuracy. It turns out that using the two combined models has higher accuracy over the quadratic polynomial model and grey model, so the combined models are effective and feasible.

Key words: quadratic polynomial model, grey model, atomic clock, lost data, residual errors, combined model