微型电脑应用

为了提高对电商直播在线评论情感分析的精度，提出一种基于稳健优化的双向编码器表征(RoBERTa)与双向长短期记忆(BiLSTM)网络的电商直播在线评论情感分析方法。通过RoBERTa对电商直播在线评论文本进行词向量表示，采用双层BiLSTM网络进行特征提取和情感识别。实验结果表明，所提出的方法在电商直播在线评论情感分析中具有一定的优越性，相较于传统的深度神经网络(DNN)、文本循环卷积神经网络(TextRCNN)、双向编码器表征(BERT)文本情感分析方法，所提出的方法的精确率、召回率、准确率、F₁值最高，分别为95.32%、95.66%、95.73%、95.46%。由此得出，所提出的方法可以提高电商直播在线评论情感分析的精度，为直播用户情感分析提供借鉴。

To improve the precision of sentiment analysis of online comments in e-commerce live streaming, a sentiment analysis method of online comments in e-commerce live streaming based on robustly optimized bidirectional encoder representation from transformers(RoBERTa) and bidirectional long short-term memory(BiLSTM) network is proposed. The online comment texts of e-commerce live streaming is represented by word vectors through RoBERTa, and the double-layer BiLSTM network is used for feature extraction and sentiment recognition. The experimental results show that the proposed method has certain superiority in sentiment analysis of online comments in e-commerce live streaming. Compared with traditional text sentiment analysis methods, such as deep neural network(DNN), text recurrent convolutional neural network(TextRCNN) and bidirectional encoder representation from transformers(BERT), the proposed method has the highest precision, recall, accuracy and F₁ value, reaching 95.32%, 95.66%, 95.73% and 95.46%, respectively. Therefore, the proposed method can improve the precision of sentiment analysis of online comments in e-commerce live streaming, which can provide reference for sentiment analysis of live streaming users.