Cross-language multi-source vertical domain knowledge graph construction method

Abstract

The invention discloses a cross-language multi-source vertical domain knowledge map construction method, which relates to the technical field of knowledge engineering. The present invention is achieved through the following technical solutions: vertical domain translation completes parallel corpus construction through content and link analysis according to input cross-language texts, domain dictionaries, domain termbases, domain materials and data; The translation model realizes automatic translation of foreign texts; domain knowledge pre-labeling training realizes active learning and labeling based on text segmentation and text clustering, completes the selection of corpus to be labeled based on analysis topics, and generates confirmed business labeling datasets; Algorithms, combining vertical domain translation data and actual scenes to complete semantic feature extraction and deep learning-based entity relationship extraction; domain knowledge fusion and disambiguation For knowledge from different sources, the network equivalent entities are merged for fusion and disambiguation to obtain cross-language multi-source knowledge Vertical domain knowledge graph.

Description

technical field [0001] The invention relates to the construction of a cross-language knowledge map in the technical field of knowledge engineering, and in particular to a method for building a cross-language multi-source vertical domain knowledge map. Background technique [0002] Knowledge Graph (Knowledge Graph) originated from related academic research fields such as semantic web and graph database. Different fields have different emphasis on knowledge graph research, such as natural language processing, knowledge engineering, machine learning, database and data management. Different research and applications. Knowledge graph construction first needs to identify available data sources, such as structured data, machine-readable open ontology or thesaurus, open linked data and open knowledge bases, industry knowledge bases and industry vertical websites, online encyclopedias (Wiki, Interactive, Baidu) and data such as text. Then, collect data effectively, such as open lin...

AI Technical Summary

Problems solved by technology

[0007] To sum up, the existing knowledge graphs mainly involve open domains, single data sources, and are mainly single-language graphs; knowledge graphs related to vertical fields are mainly concentrated in the financial field, and most of their data are structured texts; knowledge graphs have great potential in the field of public security. Potential application value, there is currently no relevant cross-language multi-source vertical domain knowledge map construction method

Images