Abstract Chinese ancient book catalogs have persisted for over two millennia, serving as ancient knowledge organization systems and cultural heritage of great academic value. This article introduces the concept of “book cluster,” a novel approach to studying catalog classification evolution, offering a meso-level perspective that bridges books and categories. Drawing from computational methods, we identify clusters of books with similar categorization trajectories, enabling more efficient and intuitive exploration of knowledge organization trends. We review existing research on catalog classification, highlighting the traditional qualitative analysis and emerging computational methods. We then present our data source and processing methods. The concept of book cluster is defined, along with the algorithm for cluster identification and feature extraction. A case study demonstrates the potential of the book cluster concept in analyzing classification changes and knowledge inheritance. Additionally, we introduce an online analytical platform for further exploration. Through computational analysis of book clusters, we gain insights into the evolution of ancient book recording, circulation, and academic contexts. Book clusters offer clues about the content of lost books. This research lays the groundwork for future iterations. By pioneering the study of knowledge organization evolution using book clusters, we enhance our understanding of Chinese history's academic culture and knowledge preservation.
Read full abstract