Automatic classification method of text syntax structure and semantic information fused text entity relationship

Abstract

The invention discloses an automatic classification method of a text syntax structure and semantic information fused text entity relationship. The method comprises the following specific implementation steps: (1) preprocessing a text set to obtain a sentence initial vector; (2) extracting relevancy between words in sentences and entities and relations, and updating sentence vectors; (3) inputtingthe sentence vector in step (2) into a bidirectional gate loop unit neural network to obtain a sentence vector fused with semantic information; (4) inputting the sentence vector in step (2) into a graph attention neural network to obtain a sentence vector fused with grammatical structure information; (5) splicing the output of step (3) and the output of step (4), and updating sentence vectors; and(6) inputting the sentence vector in step (5) into a fully connected neural network layer to obtain a feature vector of the sentence, and outputting an entity relationship classification vector through softmax transformation. The method can effectively relieve the problems that semantic features and grammatical features are difficult to express and artificial feature selection errors are large inentity relationship classification.

Description

technical field [0001] The present invention relates to the fields of text mining and deep learning, and more specifically, relates to a text entity relationship automatic classification method that integrates text grammatical structure and semantic information. Background technique [0002] Traditional entity-relationship classification methods using machine learning, such as support vector machines, need to construct a large number of lexical, syntactic and semantic features, and these features cannot be reused across domains, which greatly limits the scope of use of the method . Traditional methods of using neural networks to classify entity relationships, such as bidirectional recurrent memory neural networks and convolutional neural networks, simply use the direct mapping between words and vectors in the text, ignoring the impact of entities and relationships on the words in sentences , the effective information of entities and relationships cannot be fully utilized, s...

AI Technical Summary

Problems solved by technology

[0002] Traditional entity-relationship classification methods using machine learning, such as support vector machines, need to construct a large number of lexical, syntactic and semantic features, and these features cannot be reused across domains, which greatly limits the scope of use of the method

Traditional methods of using neural networks to classify entity relationships, such as bidirectional recurrent memory neural networks and convolutional neural networks, simply use the direct mapping between words and vectors in the text, ignoring the impact of entities and relationships on the words in sentences , the effective information of entities and relationships cannot be fully utilized, so the classification accuracy is not high; some neural network methods for entity relationship classification using sentence grammatical structure, such as the neural network method based on the shortest dependency path, ignore other components in the sentence The impact on the entity relationship itself causes the loss of lexical information other than the entity, and it is impossible to effectively obtain the positive impact of environmental words other than the entity on the entity relationship

Images