Entity matching method and system, electronic equipment and storage medium

The invention discloses an entity matching method and system, electronic equipment and a storage medium, and the method comprises the steps: combining a first entity set and a second entity set to obtain a target entity set, and constructing an undirected and unweighted graph by taking attributes of...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: BAI QIANGWEI, XUE XIAONA
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses an entity matching method and system, electronic equipment and a storage medium, and the method comprises the steps: combining a first entity set and a second entity set to obtain a target entity set, and constructing an undirected and unweighted graph by taking attributes of all entities in the target entity set as nodes; carrying out random walk sampling in the undirected and unweighted graph, converting a path obtained after sampling into a text, and utilizing the text to construct a corpus; continuously pre-training the BERT model by using a corpus; constructing a text matching model according to the pre-trained BERT model, converting the marked entity matching corpus in the corpus into a text matching corpus, and training the text matching model by using the text matching corpus; and matching to-be-matched entities by utilizing the trained text matching model. According to the method and system, the attribute context is fused into the pre-training language model, so that an entity