Discovering and linking method of knowledge graph entity based on production alias excavation

Abstract

The invention is a discovering and linking method of knowledge graph entity based on production alias excavation, which comprises the following steps that A: the entity is alias excavated and then an alias-entity mapping dictionary is produced; B: a candidate entity is produced based on the improved text edit distance; and C: the candidate entity is distinguished to attain an entity alleged item.

Description

technical field [0001] The invention relates to the field of data processing, in particular to a method for discovering and linking knowledge map entities based on production alias mining. Background technique [0002] The data in the Internet is growing exponentially, and network data mainly exists in the form of natural language. On the other hand, the rise of knowledge sharing communities such as Wikipedia has built a large-scale machine-readable knowledge base. Entity discovery and entity linking aim to identify entity referents mentioned in text and link them to corresponding entities in a given knowledge base. Through entity linking, large-scale, high-noise network data can be linked with knowledge bases, which is of great benefit to the construction of knowledge graphs. In addition, entity linking is a key step in knowledge base completion, automatic question answering, information integration, and information retrieval. . [0003] The current entity discovery and e...

AI Technical Summary

Problems solved by technology

[0005] ●Alias ​​mining mainly uses structured extraction or manually labeled alias corpus, and does not have the ability to automatically discover aliases from text, resulting in a low recall rate of words that do not exist in the alias dictionary;

[0006] ●In the calculation of the edit distance of the string, all characters are given the same weight, and the influence of the probability of word and word error on the edit distance is not considered;

[0007] ●In the discriminant model, there is a problem with the representation of context information, and the context cannot be well used for discrimination

Images