A Forecast Model based on Two-step Clustering and Random Forest
Download as PDF
DOI: 10.23977/gefhr.2019.057
Author(s)
Xiaojing Wang, Qingxia Xie, Chuantao Wang
Corresponding Author
Qingxia Xie
ABSTRACT
Demand forecasting has played an important role in inventory management of e-commerce enterprises in the era of big data. In this study, in order to improve the accuracy of forecasting, a combination model based on the Two-step Clustering and Random Forest is proposed. The Two-step Clustering Algorithm is firstly applied to clustering data series into several disjoint clusters. Then, each cluster is set as the input and output sets to construct the corresponding C-RF model. Finally, the testing set is partitioned into the corresponding cluster by the trained Two-step Clustering model, and then the prediction results are calculated based on the corresponding trained C-RF model. By comparison with the single Random Forest model, the C-RF model based on Two-step Clustering is proved to outperform the single Random Forest model.
KEYWORDS
Demand forecast, Two-Step Clustering, Random Forest, and E-commerce