计算机应用 ›› 2019, Vol. 39 ›› Issue (7): 1925-1930.DOI: 10.11772/j.issn.1001-9081.2018112340

• 人工智能 • 上一篇    下一篇

基于层次注意力机制神经网络模型的虚假评论识别

颜梦香1, 姬东鸿1, 任亚峰2   

  1. 1. 武汉大学 国家网络安全学院, 武汉 430072;
    2. 广东外语外贸大学 外语研究与语言服务协同创新中心, 广州 510420
  • 收稿日期:2018-11-26 修回日期:2019-02-16 发布日期:2019-07-15 出版日期:2019-07-10
  • 通讯作者: 任亚峰
  • 作者简介:颜梦香(1993-),女,湖北孝感人,硕士研究生,主要研究方向:自然语言处理;姬东鸿(1967-),男,河南驻马店人,教授,博士生导师,博士,主要研究方向:自然语言处理、机器学习、数据挖掘;任亚峰(1986-),男,河南焦作人,副研究员,博士,主要研究方向:自然语言处理、机器学习。
  • 基金资助:

    国家自然科学基金资助项目(61702121,61772378)。

Deceptive review detection via hierarchical neural network model with attention mechanism

YAN Mengxiang1, JI Donghong1, REN Yafeng2   

  1. 1. School of Cyber Science and Engineering, Wuhan University, Wuhan Hubei 430072, China;
    2. Collaborative Innovation Center for Language Research and Service, Guangdong University of Foreign Studies, Guangzhou Guangdong 510420, China
  • Received:2018-11-26 Revised:2019-02-16 Online:2019-07-15 Published:2019-07-10
  • Supported by:

    This work is partially supported by the National Natural Science Foundation of China (61702121, 61772378).

摘要:

针对虚假评论识别任务中传统离散模型难以捕捉到整个评论文本的全局语义信息的问题,提出了一种基于层次注意力机制的神经网络模型。首先,采用不同的神经网络模型对评论文本的篇章结构进行建模,探讨哪种神经网络模型能够获得最好的篇章表示;然后,基于用户视图和产品视图的两种注意力机制对评论文本进行建模,用户视图关注评论文本中用户的偏好,而产品视图关注评论文本中产品的特征;最后,将两个视图学习的评论表示拼接以作为预测虚假评论的最终表示。以准确率作为评估指标,在Yelp数据集上进行了实验。实验结果表明,所提出的层次注意力机制的神经网络模型表现最好,其准确率超出了传统离散模型和现有的神经网络基准模型1至4个百分点。

关键词: 注意力机制, 虚假评论, 离散特性, 神经网络, 长短期记忆网络

Abstract:

Concerning the problem that traditional discrete models fail to capture global semantic information of whole comment text in deceptive review detection, a hierarchical neural network model with attention mechanism was proposed. Firstly, different neural network models were adopted to model the structure of text, and which model was able to obtain the best semantic representation was discussed. Then, the review was modeled by two attention mechanisms respectively based on user view and product view. The user view focused on the user's preferences in comment text and the product view focused on the product feature in comment text. Finally, two representations learned from user and product views were combined as final semantic representation for deceptive review detection. The experiments were carried out on Yelp dataset with accuracy as the evaluation indicator. The experimental results show that the proposed hierarchical neural network model with attention mechanism performs the best with the accuracy higher than traditional discrete methods and existing neural benchmark models by 1 to 4 percentage points.

Key words: attention mechanism, deceptive review, discrete feature, neural network, Long Short-Term Memory (LSTM) network

中图分类号: