Text matching method based on data enhancement and graph matching network

The invention discloses a text matching method based on data enhancement and a graph matching network, which introduces dependency syntactic analysis and establishes a dependency relationship between text semantic units. And focusing the model on the core semantic unit by using a self-attention mechanism. Through a graph matching network, connection is established between any words of two text segments, full interaction is carried out through an attention mechanism, semantic focuses are better grasped, and the similarity between the texts is learned. And realizing data enhancement of the question matching data set from word and sentence granularity. An entity replacement algorithm, a synonym replacement and random insertion algorithm, a word noise enhancement algorithm and a back translation algorithm are designed, so that the diversity of a data set is improved, and the problems of insufficient corpora and the like are solved. According to the method, dependency syntactic analysis and a graph matching networ