晩唐 李商隱과 北宋 西崑派 멀리서 읽기

Abstract

In this paper, n-gram analysis, cosine similarity analysis using bi-gram, and LDA topic modeling were used to investigate the connection between Li Shang-yin’s poetry and Xikun Style’s poetry as a method of ‘Distant Reading’. By analyzing one-letter uni-gram and two-letter bi-gram, we can identify the preferred words of each poet. In particular, due to the structured format of classical Chinese poetry, which consists of a line of five or seven characters, it is easy and meaningful to segment into two-character units, so the frequency analysis of bi-gram can yield more meaningful data values for examining trends in poetry. Cosine similarity analysis using bi-gram allows us to measure the distance between poets. In this case, we found similarities in the use of words in Li Shangyin and Xikun Style poems, which are recognized as inextricably linked in literary history. LDA topic modeling is also a data analysis technique that uses bi-gram as the basis, and it allows us to examine the tendency of individual topics in Li Shag-yin’s poetry and Xikun Style poetry and the differences between them from a macro perspective.