Multi-label Classification Model for Consumer Product Defect Clues Based on BERT-GCN
Jingjing Tian, Xueqiang Lv, Xinhao Feng · 2023
Multi-label text classification is one of the important subtasks in the field of natural language processing. Multi-label classification of consumer product defect clues is the foundation for conducting multi-source data association analysis. This paper proposes a multi-label classification model for defect clues that integrates BERT (BidirectionalEncoder Representations from Transformer) and GCN (Graph Convolution Network) to address the labeling problem of consumer product defect clues. The model is compared and tested based on 10351 historical case data. The experimental results show that the accuracy of the multi-label intelligent classification model reaches higher than 65%, which can effectively improve the efficiency of defect clue analysis.