Sales Forecasting Model Based on Ensemble Learning and Its Application in Anomaly Detection
Wang Minfei, Shiyuan Tong, Pei Qi, Jianmin Ge, Ying Liu, Sun Xiao Jian · 2023
Efficient sales forecasting is essential for production companies, which is also helpful to regulate the business strategies. However, due to the complex characteristics of sales data and vulnerability to external interference, the existing prediction methods based on a single model cannot achieve satisfactory results. In this paper, in order to improve the prediction accuracy, we propose a forecasting model based on ensemble learning of both homogeneous and heterogeneous machine learning models to cope with different kinds of complexity in sales data. Results based on the Chinese cigarette sales data show that our model is competitive compared to single machine learning model and the classical time series-based predication model. Besides, the proposed forecasting model is applied to anomaly detection of sales data by evaluating the residual between the predicted value and the observed one.