Method and system used for generating knowledge graphs

The invention discloses a method and system used for generating knowledge graphs. The method includes: generating the initial knowledge graphs; judging types of graphic units; identifying association relations among named entities in the singular graphic unit by taking an initial knowledge graph framework as a constraint condition and on the basis of the types of the graphic units; extracting the association relations among newly-added named entities and the named entities in a graphic unit set to form candidate knowledge items; modifying the candidate knowledge items and updating the initial knowledge graphs according to the modified candidate knowledge items. The association relations among the multiple and abundant named entities are automatically discovered, extracted and organized from a huge number of graphic datum by the aid of computers starting from the basic knowledge graph framework and a small seed-knowledge set, a knowledge graph structure can be rapidly, effectively and elaborately presented after determination of an expert system, and a great deal of labor cost is obviously saved in the process for generating the knowledge graphs.