摘要
铁路运营维护中产生了大量非结构化的文本数据,针对这些文本信息,提出一种基于Word2Vec+MCNN的文本挖掘分类方法。首先采用Word2Vec训练故障词向量;其次丰富词向量矩阵信息,使网络模型从多方位的特征表示中学习输入句子的故障信息;最后使用多池化卷积神经网络模型作为故障分类的方法,得到更多全面的隐藏信息。通过与传统分类器以及其他类型的多池化卷积神经网络模型实验对比,得出本文的模型可以更好地达到分类效果,具有较高的分类准确率。
There are a lot of unstructured text data in railway operation and maintenance.For this text information,this article proposes a text mining classification method based on Word2Vec+MCNN.Firstly,the Word2Vec was used to train the fault word vector.Secondly,the word vector matrix information was enriched to enable the network model to learn the fault information of input sentences from the multi-dimensional feature representation.Finally,the multi-pooling convolutional neural network model was used as a fault classification method to acquire more comprehensive hidden information.Compared with the traditional classifiers and other types of multi-pooling convolutional neural network model experiments,it is concluded that the model can achieve better classification effect and higher classification accuracy.
作者
周庆华
李晓丽
ZHOU Qinghua;LI Xiaoli(School of Electronic and Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《铁道科学与工程学报》
CAS
CSCD
北大核心
2019年第11期2859-2865,共7页
Journal of Railway Science and Engineering
基金
国家自然科学基金资助项目(61763025)