基于回归分析法的农配网协议库存需求预测

Demand forecast of agricultural distribution network agreement inventory based on regression analysis

农配网协议库存采购是国家电网公司最具特色的一种采购模式,提高农配网物资需求预测准确度是实现物资采购精益化管理的重要组成部分。根据历史农配网项目建设信息、物资采购信息,提出了基于工程属性和物资小类的最小预测单元划分方法,并根据最小预测单元内样本数量,采用不同的预测分析方法,对农配网项目所需物资小类数量进行预测,进一步实现对物资金额预测。该方法可用于协助电力公司提高农配网协议库存物资需求预测准确度,同时,对需求预测工作提出了几点建议。

Agricultural distribution network is the most characteristic of the state grid corporation of a procurement model, to improve the demand for material distribution of agricultural distribution network is to achieve the accuracy of material procurement management of the important part. According to the construction information and material purchase information of the historical agricultural distribution network project, the minimum forecasting unit division method based on engineering attribute and material subclass was proposed. And according to the number of samples in the minimum forecasting unit, different forecasting methods were used to forecast the number of subordinate materials needed for the agricultural distribution network project, and the material quantity forecast was further realized. This method can be used to assist the power companies to improve the forecasting accuracy of the stock demand forecast of the agricultural distribution network agreement. At the same time, some suggestions are put forward for the demand forecasting work.