消费品多源缺陷线索信息预处理研究

摘要：

消费品召回已成为产品质量安全后市场监管最重要的措施之一。消费品缺陷线索是发现潜在缺陷的数据源，准确、及时、有效地收集与分析消费品缺陷线索是快速掌握消费品缺陷并实施召回的基础。如何基于消费品缺陷线索快速定位产品潜在缺陷是召回管理的关键，消费品具有产品类型多、故障模式杂的特点，获取缺陷线索到缺陷线索可深入分析而后初步判定潜在缺陷之间存在一定的差距，本文主要从消费品缺陷线索采集监测、标签字典、数据筛选、数据清洗等环节提出数据预处理的要求，为消费品缺陷线索分析提供参考。

Abstract:

Consumer product recall has become one of the most important measures for market supervision after product quality and safety. Consumer product defect clues are the data source for discovering potential defects. Accurate, timely and effective collection and analysis of consumer product defect clues is the basis for quickly grasping consumer product defects and implementing recall. How to quickly locate the potential defects of products based on the defect clues of consumer goods is the key to recall management. Consumer goods are characterized by multiple product types and various failure modes. There is a certain gap between obtaining the defect clues and preliminarily finding the potential defects by analyzing defect clues. This paper mainly puts forward the requirements for data preprocessing from the collection and monitoring of defect clues of consumer goods, label dictionaries, data screening, data cleaning and other links. It provides reference for the clue analysis of consumer goods defects.

基金项目：本文受中央基本业务费项目“基于多源信息融合的消费品缺陷线索智能分析关键技术研究与应用”（项目编号：282022Y-9461）资助。

作者简介：徐思红，工程师，主要研究方向为缺陷信息、舆情信息、缺陷产品召回信息等综合数据分析。张力丹，助理工程师，主要研究方向为缺陷信息采集和处理。田晶晶，高级工程师，主要研究方向为缺陷产品召回信息管理、数据挖掘应用等。齐月，助理工程师，主要研究方向为缺陷信息采集和处理。孙宁，高级工程师，主要研究方向为缺陷产品管理召回、数据挖掘应用、新闻宣传等。