Machine Learning Approach to Study the Bulgarian-Greek Dictionary of Neofit Rilsky and the Bulgarian Grammar Comparing Their Manuscripts from the First Half of the 19th Century
摘要
The article describes an approach to compare two Bulgarian language manuscripts from the first half of the 19th century – the Bulgarian Greek Dictionary of Neofit Rilsky and Bulgarian grammar. Both manuscripts are stored in the St. St. Cyril and Methodius National Library of Bulgaria and are available online. For our research, we have used the Transcribus text recognition software to scan, digitize, transcribe, and search the text of both manuscripts. Using the Transkribus Machine Learning approaches, we have created Handwritten Text Recognition models to study the text’s handwriting, script, and language. We also examined the hypothesis that both manuscripts were created by the same author.