Inter:一种面向不完整存货数据视图的异常存货识别方法

Inter:Abnormal Inventory Recognition Method for Incomplete Inventory Data View

异常存货清查与处置作为当前存货管理中较为重要的一环,大多依赖人工进行盘点与清查,个体差异可能造成部分存货数据要素缺失,从而为异常存货清查带来了挑战。本文构建了一种异常存货识别网络,同时实现了海量存货数据背景下的缺失值填补与异常存货状态识别。在大型航空制造企业中三家专业厂的真实存货数据集上进行了实验,结果表明,本文所提的Inter模型能在不完整视图场景下有效识别异常存货状态,为大型航空制造企业中的存货清查工作提供了新方法。

Abnormal inventory inventory and disposal,as a more important part of current inventory management,mostly rely on manual inventory and inventory,which causes some inventory data elements to be missing,thus bringing challenges to abnormal inventory inventory.This paper constructs an end-to-end Abnormal Inventory Recognition Encoder-Decoder Network(Inter)for incomplete view scenarios,it also realizes missing value filling and abnormal inventory status identification among the massive inventory data.The experimental analysis is carried out on the real inventory data set of three specialized factories in large-scale aviation manufacturing enterprises,which shows that the Inter model can effectively identify abnormal inventory status under the incomplete view scenario,and provides a novel idea for inventory recognition in large-scale military aviation manufacturing enterprises.