『한국문집총간』 엔그램 분석 프로그램 개발 연구

Abstract

Now the study of Korean Chinese literature is entering a new turning point. Big data processing technology allows access to research that is different from before. However, the application of research is hard to expect unless the researcher of Korean Chinese literature has an understanding of technology. Google’s "Ngram Viewer" shows how big data technology can be achieved by applying it to Korean Chinese literature. 『Korea Literature Series』 of the Korean Classical Database is a good source for the establishment of the Korean-style Ngram viewer. This research is trying to establish an Ngram viewer by reviewing problems caused by the development of the elementary level of the Ngram analysis program and looking for ways to improve it. Basically, Copus builds on the basis of analysis. However, 『Korea Literature Series』 was unfortunately designed as a searchable text concept. Therefore, it is necessary to rework the data to try to analyze big data. The difficulty of rework requires a precise algorithm to handle it because the author has mixed other works or there are data to be excluded from the analysis, such as footnotes. In addition, accurate analysis is possible only after processing the words that are synonymous with the transferor, classifying the style and processing of the title. In order to build an Ngram viewer, an Ngram analysis program is required first. Ngram analysis is possible on a syllable or syntactic word basis, but syllabic unit analysis is essential due to the nature of Korean Chinese literature data, which cannot accurately distinguish words. Through syllabic unit analysis, the currently developed program provides statistics by gram, shared gram and gram similarity Measurement for the book collection and the function of searching for similar passages and sides of the collection. However, only one-on-one comparisons are possible between the collection and the collection, and analysis of the entire collection is not yet possible. By improving the accuracy of the program by improving errors through test analysis, and adding various functions, the Korean-style Ngram Viewer is available based on this basis. The Korean-style Ngram Viewer is a research aid, and its success or failure depends on the researchers’ ability to understand the nature of the Korean Chinese literature materials and design programs that can be useful. To complete the Korean-style Ngram Viewer, the difficulties reviewed by the main office must be solved, which has limitations with individual efforts alone. We look forward to the active interest and support of the academic community.