イギリス　ニ　オケル　ミステリー　オウゴンジダイ　ノ　ジョセイサッカ　サクヒン　ノ　ヒカク　ブンセキ　Christie　Sayers　ト　Allingham

ツチムラ, ナルミ

Tsuchimura, Narumi

Graduate School of Language and Culture, Osaka University

オオサカダイガク　ダイガクイン　ゲンゴブンカ　ケンキュウカ

Random Forests

文体

特徴語

style

characteristic words

This study investigates characteristic words of works by Agatha Christie, a female mystery writer in the UK, comparing with other authors'ones. The selected authors for the comparison are Dorothy Sayers and Margery Allingham. All of them are famous female mystery writers during Britain's golden age of crime fiction. This study applied Random Forests, a machine learning method, for classifying the three authors'works and extracting characteristic words from each author's works. The accuracy of classification was a little low when short stories were included in the data for the analysis, but when only long novels were used in Random Forests, all the texts were correctly classified into three different groups with an accuracy of 100%. First, the extracted characteristic words from Christie's works are -ly adverbs. Christie might use these words to describe how the characters in her works acted or said something. Second, spoken vocabularies like yes, sure, and you are also extracted as characteristic words of Christie's works. This result should be because Christie's novels contain a lot of conversations between characters.