Scholarlink

<p>Biographical texts often fail to fully showcase their rich semantic knowledge due to traditional narrative modes and knowledge gaps between authors and readers. A multidimensional knowledge reorganization framework for biographical texts involves semantic description, fine-grained knowledge extraction, and knowledge reorganization applications. Based on ontology theory, a core conceptual model for biographical texts was established, employing GPT-4 and BERT for entity recognition. Knowledge reorganization strategies were proposed for key application scenarios and validated through case visualizations. A conceptual model for biographical texts was constructed. Significant enhancement of tag corpora was achieved through LLMs and the RoBERTa-BiLSTM-CRF model, achieving optimal fine-tuning in NER. Strategies based on temporal-spatial transformation, social network analysis, and thematic evolution were proposed, culminating in a knowledge graph centered on “Character-Works-Ideas”. Based on methods proposed by us, issues in semantic description and knowledge extraction of biographical texts have been effectively resolved, enhancing the application value of biographical resources.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-dimensional intelligent reorganization and utilization of knowledge in ‘Biographies of Chinese Thinkers’

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月5日

查看全文
  • Jiangfeng Liu,
  • Zhiyuan Liu,
  • Yu Shen,
  • Ran Zhang,
  • Ningyuan Song,
  • Jialong Liu,
  • Lei Pei

摘要

Biographical texts often fail to fully showcase their rich semantic knowledge due to traditional narrative modes and knowledge gaps between authors and readers. A multidimensional knowledge reorganization framework for biographical texts involves semantic description, fine-grained knowledge extraction, and knowledge reorganization applications. Based on ontology theory, a core conceptual model for biographical texts was established, employing GPT-4 and BERT for entity recognition. Knowledge reorganization strategies were proposed for key application scenarios and validated through case visualizations. A conceptual model for biographical texts was constructed. Significant enhancement of tag corpora was achieved through LLMs and the RoBERTa-BiLSTM-CRF model, achieving optimal fine-tuning in NER. Strategies based on temporal-spatial transformation, social network analysis, and thematic evolution were proposed, culminating in a knowledge graph centered on “Character-Works-Ideas”. Based on methods proposed by us, issues in semantic description and knowledge extraction of biographical texts have been effectively resolved, enhancing the application value of biographical resources.