基于ChineseBERT和多特征协同网络的电力设备缺陷文本分类模型

Text Classification Model of Power Equipment Defects Based on Chinese BERT and Multi Feature Cooperative Network

针对传统模型特征提取不够全面,词向量语义表达不准确等问题,提出了结合ChineseBERT和多特征协同网络的电力设备缺陷文本分类模型。采用针对汉字优化的ChineseBERT模型提取文本向量表征,提高词向量语义表示的准确性。多特征协同网络全面捕捉缺陷文本局部和上下文语义特征。软注意力机制提升模型聚焦于关键特征的能力。在真实电力设备缺陷文本数据集开展实验,结果表明该模型分类性能优于近期表现较好的深度学习模型,F1分数高达96.82%,证明了模型的有效性。

To address the problems of incomplete feature extraction and inaccurate semantic expression of word vector in traditional models,a text classification model of power equipment defects based on ChineseBERT and multi feature collaborative network is proposed.ChineseBERT model optimized for Chinese characters is used to extract text vector representation to improve the accuracy of word vector semantic representation.Multi feature collaborative network comprehensively captures the local and contextual semantic features of defective text.The soft attention mechanism improves the ability of the model to focus on key features.Experiments on real power equipment defect text data show that the classification performance of the model is better than the recent deep learning model,and the F1 score is as high as 96.82%,which proves the effectiveness of the model.