Haruki Murakami novels are popular around the world. We analyse how his novels are read in foreign countries and identify the cultural differences, using the amazon book reviews on his novels in Japan, US, UK, and Canada. We set the target novels “Colorless Tsukuru Tazaki and His Years of Pilgrimage” and “Norwegian Wood” because types and themes of the two novels are different. The reasons why we selected Haruki Murakami are first that the number of reviews is so large even in US, UK and Canada and secondly that Mr Murakami has a good command of English so the translation has done quite well owing to his sense of English. We think that the good translation keeps the essence of the novels as the original Japanese ones.The core technology of this analysis is text mining. The morphological analysis tool for Japanese named “MeCab” and one for English named “TreeTagger” are used. In addition, by using “word2vec”, vector representations of review words have been conducted in the text mining. The word2vec offers the high measurement quality of representations in a word similarity task. We would like to analyse the change of the word representations in the vector space among the four countries’ reviews. There are many reviewers for the text data sets. We would like to analyse whether the review tendency is different up to the country or not, and the individual authors’ opinions could be divided into the same kind of topic classes or not.
Yukari Shirota, Gakushuin University, Japan
Takako Hashimoto, Chiba University of Commerce, Japan
Yuriko Yano, Gakushuin University, Japan
Stream: Cultural Studies
This paper is part of the ACCS2017 Conference Proceedings (View)
View / Download the full paper in a new tab/window