Sung, Hsin-HsuanHsin-HsuanSungChen, Jou-AnJou-AnChenJIEH HSIANG2020-05-042020-05-042019https://www.scopus.com/inward/record.uri?eid=2-s2.0-85072867192&doi=10.1007%2f978-3-030-30760-8_24&partnerID=40&md5=9c7c1b68c8d4045394c2f40a133bb963In this paper we describe a method for finding relevant texts in historical documents. Instead of using keyword search, it computes the level of relevance between texts. Using a subset of documents that are known to be relevant, our method computes the degree of relevance between the pre-selected texts and the remaining texts in the document set. The potential texts of interest obtained from the computation are listed by their ranks and given to historians to confirm their relevance. The marked results are then fed back to the system to start the next iteration. We applied our method to find Taiwan-related entries in the Veritable Records of Qing, and discovered an additional 988 related entries which were not included in the volume of Veritable Records of Qing – Taiwan Collection (which contains 2989 entries). © Springer Nature Switzerland AG 2019.Computation theory; History; Iterative methods; Search engines; Degree of relevance; Historical documents; Keyword search; Relevance feedback; Digital librariesconference paper10.1007/978-3-030-30760-8_242-s2.0-85072867192https://doi.org/10.1007/978-3-030-30760-8_24