텍스트마이닝을 활용한 ‘호랑이’ 설화의 전승 시기별 지역 간 유사성 분석

Abstract

The present study analyzed the similarities in the transmission trends among 8 regions by dividing the “Tiger” folktales listed in Comprehensive Korean Oral Literature and Complementary Edition of Comprehensive Korean Oral Literature into each compilation. For this purpose, text mining techniques were utilized, involving the following process : data collection → modification and supplementation of regional information and titles of tiger folktales → analysis of regional folktales → visualization. First, we collected title data from 26,542 traditional tales and preprocessed them by identifying regions where administrative divisions were not specified at the province level. Additionally, we replaced titles labeled as ‘horaengi’ (호랭이) or ‘범’ (beom) with ‘tiger’ (호랑이), and for tales involving tiger stories where ‘tiger’ did not appear in the title, we added the word ‘tiger.’ After completing this preprocessing step, we extracted the tiger folktales. Next, we classified these stories by region and divided them according to the Comprehensive Korean Oral Literature and the Complementary Edition of Comprehensive Korean Oral Literature. From this data, we created a corpus consisting only of titles and analyzed the frequency of nouns to extract data. To enable accurate comparison between regions, we normalized the frequency of nouns. Then, using the calculated normalization values, we computed cosine similarity values to compare them between each region, thereby analyzing the similarity in the transmission trends of tiger folktales among different regions according to transmission periods. As a result, at the time of the survey of Comprehensive Korean Oral Literature (1979~1985), it was confirmed that there were many areas where the story of tigers killing people was told with interest, and that local tiger tales were actively passed down, such as the legend of ‘Hwangpaldo’(황팔도) in South Chungcheong. At the same time, in the Chungbuk and Jeonbuk regions, where many folktales of geomancy are passed down, tiger tales have also been spotted showing such characteristics. At this time, the similarity between regions of the tiger tales passed down was found to be quite low in most regions except for North Gyeongsang-South Gyeongsang and North Gyeongsang-South Jeolla. On the other hand, in Complementary Edition of Comprehensive Korean Oral Literature (2008~2018), it was found that ‘Tiger and Dried Persimmon’ were collected in a large proportion in many regions, and in Gyeonggi, Gangwon, and Chungnam regions, the type of ‘Brother and Sister who Became the Sun and Moon’ was also passed down particularly actively. These three regions were found to be regions with a relatively high degree of similarity in the trend of tradition, which is different from the trend of tradition in the survey of Comprehensive Korean Oral Literature. In addition, Gyeongnam, Jeonnam, and North Gyeongsang Province showed more similarity over time, and Jeonbuk and Chungnam emerged as regions with new similarities to Gyeongnam.