基于深度学习的电商用户评论情感极性分析

Analysis on Affective Polarity of E-commerce Users’Comments Based on Deep LearningWU Shu-fan

为对电商平台中的用户评论进行情感极性分析,先对这些电商评论文本进行预处理,再利用经过训练优化了参数的BERTCNN模型进行分析。结果表明,BERTCNN模型虽然用时较长,但准确率、召回率和F都优于其他模型,而且BERTCNN模型训练过程稳定,收敛后损失较小,因此,在情感分析中BERTCNN模型效果最好。

In order to analyze the emotional polarity of user reviews in e-commerce platform,these e-commerce comment texts are preprocessed,and then analyzed by using BERT-CNN model,which has been trained and optimized,but the accuracy,recall and F of BERT-CNN model are better than other models,and the training process of BERT-CNN model is stable with less loss after convergence.