Prediction of Material Demand for Power Engineering Based on Data Space
Accurate material demand prediction assists the material management department in developing accurate procurement strategies to reduce costs.However,there are many factors that affect the demand of power engi-neering material,and these data belong to different departments.The lack of collaboration between these departments has formed a serious"Data Silo"phenomenon.At the same time,the sample size is small and the influencing factors considered are not comprehensive when enterprises predict the total annual demand for power engineering materials,which makes the results underperform.This paper builds a power engineering material Data space integrating the value of data from multiple departments,explores important information intext-based data through Natural Language Processing,and applies tree models to explore the potential connection between multi-source heterogeneous data from an engineering project perspective to predict the material demand of power engineering accurately.The research results show that the prediction accuracy of the proposed model in this paper is improved by 30percent,which is more stable than the bench model and its current practice.It provides effective decision-making support for the case company's material management.
Demand prediction of power engineering materialData spaceTree models
万方数据
维普
