End-to-end entity relation joint extraction method and system based on relation decomposition

The invention discloses an end-to-end entity relationship joint extraction method based on relationship decomposition, which is characterized by comprising the following steps of: data preprocessing: converting triples of entities and relationships labeled in a training set into a vector form accord...

Full description

Saved in:
Bibliographic Details
Main Authors DU KUNPENG, ZHANG XUAN, YUAN ZIHAO, NONG QIONG, GAO CHEN, MA QIUYING
Format Patent
LanguageChinese
English
Published 31.05.2022
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses an end-to-end entity relationship joint extraction method based on relationship decomposition, which is characterized by comprising the following steps of: data preprocessing: converting triples of entities and relationships labeled in a training set into a vector form according to a dictionary in a BERT model; model training: carrying out relation classification according to text vectors output by the BERT model, and then fusing relation features with sentence features to carry out head and tail entity recognition: result decoding: decoding entity tags recognized under different relation categories, and combining the decoded entity tags with relations to obtain entity relation triples existing in sentences. According to the method, the sentence features under different relations are modeled, so that the extraction problem of the overlapped triples in the sentences can be effectively solved, the performance of entity relation joint extraction is improved, and the method has good practi
Bibliography:Application Number: CN202210166252