关键词: GloVe, 门控循环单元（GRU）, 文本分类

Abstract:

Text classification has a wide range of applications, and the research of its classification algorithm has been concerned. However, traditional text classification algorithms generally have some problems, such as too high dimension of text feature vectorization, not considering the semantic relationship between keywords, too many training parameters, which will affect the performance of classification accuracy and so on. In order to solve these problems, this paper proposes a text classification algorithm which combines word vectorization and GRU. First, it preprocesses the text. Then it extracts features through GloVe to contain as much semantic and grammatical information as possible, while reducing the vector space dimension. Finally, it uses GRU neural network model for training to retain the semantic association between long-distance words in the long text to the greatest extent. The experimental results show that the algorithm can improve the performance of text classification.

Key words: GloVe, Gated Recurrent Unit（GRU）, text classification