Literature author name duplication disambiguation method and literature author name duplication disambiguation construction system

Abstract

The invention discloses a literature author name duplication disambiguation method and a literature author name duplication disambiguation construction system. The literature author name duplication disambiguation method comprises the following steps of 1, reading literature data and scholarship data in a database; 2, training and predicting a document vector of each document by using a Word2Vec model; 3, constructing a to-be-disambiguated author partner relationship network graph, and calculating node similarity and clustering; and 4, obtaining document vectors of documents in document clusters clustered by the partner relation graph, and calculating similarity and clustering among the document clusters. The invention can ensure that the disambiguation result has relatively high accuracyand recall rate level, and is suitable for multi-language and multi-literature types of Chinese literature, English literature, patents and the like.

Description

technical field [0001] The invention belongs to the technical field of document processing, in particular to a method for disambiguation of duplicate names of document authors. Background technique [0002] With the rapid development of science and technology and the continuous integration of information, when dealing with informatization issues, especially when dealing with flexible and diverse natural language data, the phenomenon of duplicate names widely existing in the real world will greatly affect the retrieval and processing of data , thus resulting in the technique of named entity disambiguation, which studies how to match ambiguous entity references with correct entities in a knowledge base. Author disambiguation belongs to named entity disambiguation. In the real world, different people may have the same name. In many applications such as scientific literature management and information integration, people’s names are used as identifiers for retrieving information...

AI Technical Summary

Problems solved by technology

[0003] The clustering technology can be used to complete the task of disambiguation of the author’s duplicate name. Most of the existing methods are based on the information contained in the literature, mainly including the method based on feature distinction, the method based on graph segmentation and the classification based on network resources. Although these methods can disambiguate duplicate names, the division methods based only on text features or graph relationships cannot fully utilize the rich information contained in the literature, and it is difficult to ensure that the disambiguation results have a high level of accuracy and recall. Moreover, the existing duplicate name disambiguation method cannot be applied to multi-language and multi-document types such as Chinese documents, English documents and patents.

Images