Safe inventory prediction method and system based on future sales volume

The invention belongs to the technical field of intelligent data processing, and mainly relates to a safe inventory prediction method and system based on future sales volume, and the method comprises the steps: obtaining historical sales volume data, and dividing the historical sales volume data into a training set and a test set; the historical sales volume data are preprocessed, and preprocessing comprises the steps that feature extraction is carried out, and missing data are filled up through a random forest algorithm; meanwhile, abnormal values in the historical sales volume data are identified, and an XGBoost model is used for correction; establishing a self-adaptive ARIMA model, and inputting the preprocessed high-effectiveness historical data into the self-adaptive ARIMA model for training; and sales volume prediction data is obtained through an adaptive ARIMA model, and warehouse replenishment transfer is performed according to the high-effectiveness historical data and the sales volume prediction dat