A method and storage device for identifying hyponym relations in knowledge graphs based on pattern expansion and Bert classification

Abstract

The invention relates to the technical field of text recognition, in particular to a method and a storage device for identifying the hyponym relationship of a knowledge map based on pattern expansion and BERT classification. The method for identifying the hyponym relationship of a knowledge map based on pattern expansion and BERT classification comprises the steps of: extracting potential hyponym relationship pairs from external data based on pattern expansion, and forming a hyponym relationship model training in combination with a preset resource library Seed corpus; obtain the data set of hyponymia to be predicted, predict the data set to be predicted based on the BERT-Attention-Bi-LSTM model, and obtain the prediction result of the hyponymia; further process the prediction result of the hyponymia through preset rules to obtain The final hyponymy prediction results. This method greatly reduces the complexity and manual input of rule compilation, and at the same time, compared with another mainstream statistical-based upper and lower concept recognition method, it has stronger realizability and can provide technical support for the construction of various professional knowledge graphs. support.

Description

technical field [0001] The invention relates to the technical field of text recognition, in particular to a method and a storage device for identifying the upper and lower relations of a knowledge graph based on pattern expansion and BERT classification. Background technique [0002] With the rapid development of Internet software and hardware related technologies, people gradually enter the intelligent age from the information age. As an important carrier that carries massive underlying knowledge and supports upper-level intelligent applications, knowledge graph plays an extremely important role in the era of intelligence. Limited by the huge difference between unstructured text and structured knowledge, there are still many challenges in automatically constructing knowledge graphs and using knowledge graphs to support upper-level applications. Therefore, it is necessary to conduct research on the construction of knowledge graph and its core application of natural language...

AI Technical Summary

Problems solved by technology

[0013] To this end, it is necessary to provide a method for identifying the hyponymy relationship of knowledge graphs based on pattern expansion and BERT classification, which is used to solve the problem that the existing knowledge graph hyponymy relationship identification methods rely heavily on professional knowledge and rule-making capabilities, requiring investment A lot of labor, relatively high demand for prior knowledge bases, etc.

Images