基于TextCNN-Bert融合模型的不良信息识别技术

Bad information identification technology based on TextCNN-Bert fusion model

敏感领域的不良信息具有极强的迷惑性和欺骗性,腐蚀人们的思想,影响人们的价值观和判断能力,危害社会安全,研究敏感领域不良信息的识别技术具有深远意义。通用的识别技术忽略了背景知识和隐喻问题,直接应用于敏感领域不良信息识别效果较差。提出一种基于TextCNN-Bert的融合模型,通过敏感领域主题识别和情感隐喻识别,实现对敏感领域不良信息的文本识别。实验结果表明,该模型在准确率、F 1评分等指标方面取得了良好的结果,相较于现有模型有显著提高。

The bad information in sensitive areas is extremely confusing and deceptive,corrodes people′s thinking,affects people′s values and judgment,and endangers social security.Research on the identification technology of bad information in sensitive areas has far-reaching significance.The general recognition technology ignores background knowledge and metaphor problems,and the effect of direct application to sensitive areas is poor in the recognition of bad information.This paper proposes a fusion model based on TextCNN-Bert,which realizes the text recognition of bad information in sensitive areas through topic recognition and emotional metaphor recognition.The experimental results show that the proposed model achieves good results in terms of accuracy,F 1 score and other indicators,which are significantly improved compared with the existing models.