Abstract:

With the explosive growth of cultural heritage information resources, efficient extraction methods of unstructured data affects the spread of Chinese traditional culture. This article takes the text of cultural heritage information resources as the research object, proposes to classify them by sources, select keyword extraction methods according to different knowledge distribution characteristics, and retrieve cultural heritage entities and relationships in the knowledge graph after obtaining keywords to complete the path of knowledge extraction. The experimental results show that the categorized keyword extraction method established by the research can improve more than 50% compared with other methods, and can better extract the cultural heritage knowledge in unstructured data.

Key words: knowledge extraction, 　keyword extraction, 　cultural heritage, 　TF-IDF, 　LDA