Ensemble learning-based cyclic serial number usage amount early warning method and ensemble learning-based cyclic serial number usage amount early warning system

The invention provides an ensemble learning-based cyclic serial number usage amount early warning method and an ensemble learning-based cyclic serial number usage amount early warning system. The method comprises the steps of obtaining serial number data in a preset time period before current time; predicting the serial number according to the serial number data through a preset ensemble learningmodel to obtain the serial number usage amount in a future time period, wherein the ensemble learning model is used for predicting the serial number according to multiple prediction algorithms and theweight of each prediction algorithm; and obtaining early warning information about whether the serial number is used up or not in a future time period according to the serial number usage amount. According to the invention, accurate monitoring and early warning of the cyclic serial number under the conditions of suddenly increased usage amount and different scenes can be satisfied. 本发明提供了一种基于集成学习的循环序列号使用量预警方法及系统，所述方法包括：得到