|
|||||
|
|
| 多类型注意力下参数自适应的多标签文本分类 | |
| 引用本文： | 李智强,过弋,王志宏.多类型注意力下参数自适应的多标签文本分类[J].中文信息学报,2022,36(10):116-125. |
| 作者姓名： | 李智强 过弋 王志宏 |
| 作者单位： | 1.华东理工大学 信息科学与工程学院,上海 200237; 2.大数据流通与交易技术国家工程实验室 商业智能与可视化技术研究中心,上海 200436; 3.上海大数据与互联网受众工程技术研究中心,上海 200072 |
| 基金项目： | 国家重点研发计划(2018YFC0807105);上海市科学技术委员会科研计划项目(22DZ1204903,2251104800) |
| 摘 要： | 多标签文本分类是指从一个极大的标签集合中为每个文档分配最相关的多个标签。该文提出一种多类型注意力机制下参数自适应模型(Parameter Adaptive Model under Multi-strategy Attention Mechanism,MSAPA)对文档进行建模和分类。MSAPA模型主要包括两部分: 第一部分采用多类型注意力机制分别提取融合自注意力机制的全局关键词特征和局部关键词特征及融合标签注意力机制的全局关键词特征和局部关键词特征;第二部分采用多参数自适应策略为多类型注意力机制动态分配不同的权重,从而学习到更优的文本表示,提升分类的准确率。在AAPD和RCV1两个基准数据集上的大量实验证明了MSAPA模型的优越性。 |
| 关 键 词： | 多类型注意力机制 参数自适应 多标签文本分类 |
| 收稿时间： | 2020-12-24 |
Parameter Adaptive Model Under Multi-Type Attention for Multi-label Text Classification |
|
| LI Zhiqiang,GUO Yi,WANG Zhihong.Parameter Adaptive Model Under Multi-Type Attention for Multi-label Text Classification[J].Journal of Chinese Information Processing,2022,36(10):116-125. | |
| Authors: | LI Zhiqiang GUO Yi WANG Zhihong |
| Affiliation: | 1.Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai 200237, China;2.Business Intelligence and Visualization Research Center, National Engineering Laboratory for Big Data Distribution and Exchange Technologies, Shanghai 200436, China;3.Shanghai Engineering Research Center of Big Data & Internet Audience, Shanghai 200072, China |
| Abstract: | Multi-label text classification assigns the most relevant multiple labels to each document from a huge label set. This paper proposes a parameter adaptive model under a multi-strategy attention mechanism (MSAPA) for multi-label text classification. The MSAPA model first uses multiple types of attention mechanisms to extract global and local keyword features with self-attention mechanism and label attention mechanism, respectively. Then it adopts a multi-parameter adaptive strategy to dynamically assign weights to multiple types of attention mechanisms, so as to learn a better text representation for classification. Experiments on the two benchmark data sets of AAPD and RCV1 validate the superiority of the MSAPA model. |
| Keywords: | multi-type attention mechanism parameter adaptation multi-label text classification |
| 点击此处可从《中文信息学报》浏览原始摘要信息 | |
| 点击此处可从《中文信息学报》下载全文 | |