A Time Series Combined Forecasting Model Based on Prophet-LGBM

With the advent of the era of traffic, forecasts based on time series are improving the efficiency of the entire society in many ways, such as website traffic trends, price trends and so on. At present, the single model is a low accuracy rate prediction method and cannot well process the complex characteristics of time series. Therefore, this paper proposes a Prophet-LGBM combination model. The combination model makes full use of the advantages of Prophet algorithm in processing time series without feature engineering and LGBM algorithm can add other types of features and lightweight. The method of rolling prediction and weighted sum is used to improve the prediction accuracy of time series. In the experiments of this paper, we design and implement the comparative experiments between the Prophet-LGBM combination model and other single models, and verify them on multiple data sets. The experimental results show that the Prophet-LGBM combination model proposed in this paper has strong applicability and high accuracy in time series prediction.