한시 유사도 분석 방안 연구

Abstract

This paper aims to develop algorithms for Chinese poetry similarity analysis. Chinese poetry has the advantage of condensing and expressing because of the limitation of the number of characters, but new expressions are difficult in situations where there are many built works. So quote(用事) and magic transformation(點化) became a technique. However, excessive use is plagiarism. At the heart of the analysis is proving the author's uniqueness. This requires confirmation of differences from other cities. The city's evaluation may vary depending on whether there are other similar poems or, if so, to what extent. The measurement of similarity is not just for identifying similarities, but for identifying influences or common characteristics through this. Poetry consists of a surface structure and a hidden structure, and these two structures work to reach the reader. Polymorphism occurs in this process. The meaning may change by one character in the same seven-letter first pitch, but if six letters are the same, it is unlikely to be a coincidence. Tracking the history of first-time users and utilization can infer their impact relationships. Analysis that finds similar verses in many poems cannot be solved without the help of a computer program. The basis of the program is the development of algorithms. In this paper, we present a development method of algorithms for these analyses. In order to measure similarity, an analysis of Sheer is first needed. It is very difficult to set the standard for shears in Chinese poetry, which is usually used seven characters are usually used. Therefore, we determine the author's poem by analyzing n-grams and identifying the frequency of use. And except for the words that are frequently used compared to other literary works, it becomes a poem or first pitch containing the writer's personality. The reason why it is difficult to measure similarity is that only the same poem or verse with the same arrangement of Chinese characters should not be judged by the same poem. In order to avoid allegations of plagiarism, Chinese characters may be reordered or replaced with similar Chinese characters. Thus, the same meaning of the word is an important subject to consideration in measuring similarity. The development of algorithms requires identifying the variant types of shears. It is necessary to collect various cases, such as simply changing five words to seven words or dividing the first pitch. Accurate similarity measurements can be made once similar shears are classified and transformed structures are identified. Using these methods will help to identify classifications or mutual influence relationships according to literary trends in works.