蓝雯飞,徐 蔚,王 涛.基于卷积神经网络的中文新闻文本分类[J].中南民族大学学报自然科学版,2018,(1):138-143
基于卷积神经网络的中文新闻文本分类
Text Classification of Chinese News Based on Convolutional Neural Network
  
DOI:
中文关键词: 自然语言处理  深度学习  卷积神经网络  注意力机制  文本分类
英文关键词: natural language processing  deep learning  Convolutional Neural Network  attention mechanism  text classification
基金项目:国家自然科学基金资助项目(61379059)
作者单位
蓝雯飞,徐 蔚,王 涛 中南民族大学 计算机科学学院武汉 430074 
摘要点击次数: 177
全文下载次数: 236
中文摘要:
      经典的卷积神经网络文本分类模型仅仅着眼于全局特征,没有考虑到局部特征.为了解决此问题,引入了注意力机制,用于提取文本中的关键词, 把全局特征与局部特征综合在一起,使得文本的特征表达更加丰富.实验结果表明:卷积神经网络分类模型比传统的机器学习方法分类效果更好, 而引入注意力机制后的卷积神经网络模 型相比于经典的文本分类模型,分类效果也有了一定程度的提高.
英文摘要:
      The classical convolutional neural network text classification model only focuses on the global features, without taking into account the local features. To solve this problem, the attention mechanism is introduced to extract keywords from the text. In this way, the global features and local features are combined together, which makes the feature representation of the text richer. Experimental results show that the text categorization model of Convolutional Neural Network is better than the traditional machine learning methods. Compared with the classical text classification model, after introducing attention mechanism, the performance of Convolutional Neural Network classification model has been improved
查看全文   查看/发表评论  下载PDF阅读器
关闭