Digital and Intelligent Processing of Tangut Script: Constructing Knowledge Graph of Tangut Character Components

本文研究了西夏文字的数智化问题，从构建西夏文字组件的知识图谱视角拓宽了西夏学的研究范围。本文首先给出了西夏文字的组件分解方法，然后研究了西夏文字组件的知识图谱构建方法，接着讨论了组件知识图谱的智能算法，最后搭建了西夏文字组件知识图谱的数智化平台。西夏文字偏旁部首组件通过各种排列组合后的可能结果，涵盖了超过20万个潜在的西夏文字。通过这种方式，我们可以探索未出现的西夏文字的更多可能性。本文提供的西夏文字结构和组成的系统化研究，使西夏文拆分有了方便的数智化工具，为进一步的文字解码和语言学研究奠定数智化的基础。西夏文字的组件的知识图谱，由于其组件的灵活性及计算机介入后的智能性，也许为挖掘西夏文字这个文化遗产提供更多机会。This paper explores the digital and intelligent processing of Tangut script, expanding the scope of Tangut studies from the perspective of constructing a knowledge graph of Tangut character components. We first propose a method for decomposing Tangut characters into components, followed by a systematic approach to constructing a knowledge graph based on these components. We then discuss intelligent algorithms applicable to this component-based knowledge graph and introduce a digital platform for managing and exploring the graph. The combinatorial possibilities of radicals and components in Tangut script potentially cover over 200,000 unique character forms, enabling the discovery of previously undocumented Tangut characters. This systematic study of the structure and composition of Tangut script offers a practical digital tool for character decomposition and lays a foundation for further decoding and linguistic research. The component-based knowledge graph, enhanced by its structural flexibility and computational intelligence, opens new possibilities for uncovering and revitalizing this cultural heritage.