A Time Series Prediction Algorithm Based on BiLSTM and Prophet Hybrid Model

This paper introduce BiLSTM-Prophet, a successor to Facebook Prophet, which set an industry standard for explainable and user-friendly forecasting frameworks. with the increase of time series complexity, it is still a challenging task to improve the prediction accuracy of time series. We view Prophet as a precursor to such a solution. However, for some complex time series, the Prophet is too idealistic, hybrid solutions are needed to bridge the gap between classical methods and neural network models. Experiments with real data in enterprise BI business demonstrate that the prediction accuracy of the BiLSTM-Prophet proposed in this paper is significantly better than that of other traditional models.