Knowledge association network key node discovery method based on topology and literature characteristics

Abstract

The invention discloses a knowledge association network key node discovery method based on topology and literature characteristics. According to the method, a literature library related to a specified keyword in a certain scientific research field and an entity name library in the research field are utilized to obtain a matching relationship between literatures and entity names. And the method also includes establishing a knowledge association network by utilizing the matching relationship, and calculating topological characteristics and literature characteristics of each node; and giving an optimal feature set of the research field on a known associated entity set by applying a Q sequence statistic thought and combining a greedy selection strategy, and applying the optimal feature set to a literature library related to other keywords in the research field to obtain a key entity. According to the invention, the knowledge association network can be extracted for the literature library related to a certain keyword in the specified research field, the feature set is optimized to obtain the optimal feature set, the optimal feature set has high robustness and mobility in the research field, and the invention is suitable for key entity discovery of the literature library which is not studied thoroughly.

Description

technical field [0001] The invention belongs to the technical field of discovering key nodes in complex networks, and in particular relates to a method for discovering key nodes in a knowledge-associated network based on topology and document features. Background technique [0002] Facing a large amount of literature knowledge that has not yet been clarified in a certain field, it is of great significance to discover the key entities in it for the literature pre-research. How to extract the core feature set from the well-studied literature database in this field is a difficult problem. For example, key entities in some fields of scientific research are more likely to appear in earlier published documents, or in more published documents, or appear on the core nodes in the knowledge association network topology. The research idea of ​​the present invention is to guide unstudied literature databases to discover key entities by extracting important documents and topological fea...

AI Technical Summary

Problems solved by technology

However, the research on how to combine the feature directions of multiple different sources is still insufficient. The general weighting method is not robust enough, and the reliability of the results is also worrying. The feature set is output in the same way, and it is applied to another literature library in the same field where the research is not yet clear to discover key entities.

Images