Text Semantic Representation Based on Knowledge Graph Correction

Text semantic representation is the foundation of natural language processing tasks. Most existing text representation methods are based on the bag-of-words model to represent text composition. However, they ignore the textual semantics of phrases, which affects the performance of downstream natural language processing tasks. In this paper, we propose a Text Semantic Representation Method Based on Knowledge Graph Correction, called TRKGC. Firstly, we adopt PCFG and CKY methods to construct the parsing tree in units of sentences. Secondly, we traverse the parse tree according to the idea of hierarchical traversal to get the phrase representation of texts. Finally, we utilize the knowledge graph to correct text representation and improve text representation semantics. Extensive experiments on text classification and text clustering demonstrate that TRKGC outperforms baseline methods in terms of text semantic representation.