Researchers have introduced a computational framework named IFST to systematically identify key historical figures within intricate Chinese historiographical texts. Developed by Shiyuan Zhou, Hongbing Liu, Junhao Wu, and Jiale Wang, the system addresses the narrative complexity of extensive literary corpora, where dense historical plotting frequently impedes a reader’s ability to grasp essential narrative content. Because key figures act as primary catalysts for plot progression, identifying them computationally offers a structured alternative to conventional close-reading approaches, which have traditionally dominated literary analysis with limited computational assistance.
The method, formally titled Identifying key Figures in Chinese historiographical texts through figure-centric Social networks and self-seeded TrustRank, operates in distinct analytical stages. First, IFST constructs a specialized figure-centric social network from the target corpus. Within this network model, individual figures serve as nodes, while connections between them represent direct co-appearances occurring within sentences throughout the historiographical text. The weights assigned to these connecting edges reflect the exact frequency of co-occurrences, capturing the intensity of interpersonal coupling between figures across the document.
In the second stage of the process, the researchers implemented a self-seeded TrustRank algorithm to evaluate the network structure. This algorithm calculates an individual importance score for each figure based on their position and relational coupling within the figure-centric social network. After scoring, the system ranks all figures in descending order according to their calculated importance values. The historical figures appearing at the apex of this ranked list are subsequently designated as the core key figures of the narrative.
To evaluate the framework’s effectiveness, the team tested IFST on Records of the Three Kingdoms, a foundational Chinese historiographical work recognized for its intricate military and political narrative. Empirical analysis demonstrated that the IFST algorithm significantly outperformed baseline approaches in correctly pinpointing central figures. Furthermore, the figures identified at the apex of the ranking exhibited strong alignment with human interpretive judgments, demonstrating the model’s validity as both an analytical research instrument and a reading aid for complex literary texts.
Subject of Research: Computational identification of key historical figures in Chinese historiographical texts
Article Title: Identifying key figures in Chinese historiographical texts via figure-centric social networks and self-seeded TrustRank
Article References: Zhou, S., Liu, H., Wu, J., & Wang, J. (2026). Identifying key figures in Chinese historiographical texts via figure-centric social networks and self-seeded TrustRank. PLOS One, 21(11), e0362803. https://doi.org/10.1371/journal.pone.0362803
Image Credits: AI Generated
DOI: 10.1371/journal.pone.0362803
Keywords: Chinese historiography, Records of the Three Kingdoms, key figure identification, figure-centric social networks, self-seeded TrustRank, computational literary analysis, interpersonal coupling, textual co-appearance, narrative structure, literary corpora, network analysis, historiographical texts, importance scoring, reading aids, plot development catalysts, human interpretive judgments, sentence co-occurrence
Cite Scienmag News
Glenn Wilkins. (November 20, 2026). Computational Network Analysis Identifies Key Figures in Classical Chinese Texts. Scienmag. https://scienmag.com/computational-network-analysis-identifies-key-figures-in-classical-chinese-texts/
Glenn Wilkins. "Computational Network Analysis Identifies Key Figures in Classical Chinese Texts." Scienmag, 20 November 2026, https://scienmag.com/computational-network-analysis-identifies-key-figures-in-classical-chinese-texts/. Accessed 20 November 2026.
Glenn Wilkins. "Computational Network Analysis Identifies Key Figures in Classical Chinese Texts." Scienmag. November 20, 2026. https://scienmag.com/computational-network-analysis-identifies-key-figures-in-classical-chinese-texts/








