Study of the monitored method of atomic clock data exception based on the model of dynamic neural network time series-NAR

Abstract

Atomic clock data of collected in laboratory has the characteristic of time series, so the atomic clock data forecasting algorithm based on the model of dynamic neural network time series-NAR (Nonparametric regression) is proposed according to the study of dynamic neural network algorithm, And the monitored method of atomic clock data exception is designed according to this algorithm. The monitored method was verified by cesium atomic clock data, results show that the proposed method in this paper is feasible, it can be monitored in real time and effectively that the possible phase jump of atomic clock correlation data.