Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (16): 108-114.DOI: 10.3778/j.issn.1002-8331.1805-0146

Previous Articles     Next Articles

Commentary Sentiment Classification Model Combining Product Features

YU Tao, LUO Ke   

  1. 1.School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha 410114, China
    2.Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation, Changsha University of Science and Technology, Changsha 410114, China
  • Online:2019-08-15 Published:2019-08-13



  1. 1.长沙理工大学 计算机与通信工程学院,长沙 410114
    2.长沙理工大学 综合交通运输大数据智能处理湖南省重点实验室,长沙 410114

Abstract: Combining the commentary information of different products to construct intelligent sentiment classifier, this paper proposes an online product commentary sentiment analysis model combining product features:PWCNN(Product Weight Convolution Neural Network). The model firstly combines word vectors of product word features by vector multiplication, and the result is taken as input of the model. Then, according to the importance of sentences, a pooled weighted convolutional neural network is used to learn the document-level representation of the reviews. In order to prevent over-fitting and improve the generalization ability, a dropout strategy is adopted at the output layer. The experimental results show that the PWCNN model achieves the best results in terms of average accuracy and [F1] value, and the training speed is improved.

Key words: sentiment classification, convolutional neural network, product word vector, weighted pooling layer, dropout algorithm

摘要: 结合不同产品的评论词信息来构建智能化的情感分类器,提出一种结合产品特征的在线商品评论情感分类模型PWCNN(Product Weight Convolution Neural Network)。模型首先进行产品词特征的词向量训练,将评论文本以及产品信息进行向量乘法组合,结果作为模型输入。然后根据句子的重要性,采用池化加权的卷积神经网络来学习评论的文档级表示。为了防止过拟合且提高泛化能力,在输出层采用dropout策略。实验结果表明,PWCNN模型在平均准确率和[F1]值等指标上取得最好结果,且提高了模型训练速度。

关键词: 情感分类, 卷积神经网络, 产品词向量, 加权池化层, dropout算法