論文 pp 要旨キーワード N 1. はじめに ; 2000; 2004 authorship attribution Brinegar 1963; 2004; Uzuner and Katz 2005; Mingzhe and Ming

Size: px

Start display at page:

Download "論文 pp 要旨キーワード N 1. はじめに ; 2000; 2004 authorship attribution Brinegar 1963; 2004; Uzuner and Katz 2005; Mingzhe and Ming"

しげじろうやがい
5 years ago
Views:

1 計量国語学アーカイブ ID 種別タイトル Title 著者 Author 掲載号発行日 KK 論文開始ページ 187 終了ページ 210 著作権者中古日記文学の計量国語学的分析と異本間の関係性の客観分析和泉式部日記と更級日記を題材に Metric Japanese study on diary literature in the Heian period and objective measurement on relationship between variants: Analyses on Izumishikibu nikki and Sarashina nikki 太刀岡勇気 TACHIOKA Yuuki 29 巻 6 号 2014 年 9 月 20 日計量国語学会

2 論文 pp 要旨キーワード N 1. はじめに ; 2000; 2004 authorship attribution Brinegar 1963; 2004; Uzuner and Katz 2005; Mingzhe and Minghu 2012 Uzuner and Katz 2005 Mingzhe and Minghu ; Grimmer and Stewart b n-gram 187

3 計量分析手法

4 異本を生成するためのデータベース 1 d{[t1] 1[t2] 2@ } 1 [t1] AA [t2] BB CC [t1] t1 2.2 n-gram 分析 n-gram n-gram n-gram - -* 3-gram gram 2 3-gram 1-gram a; 2001; 2012 n-gram 0 0 n-gram n-gram n-gram

5 1 3-gram 2 3-gram 3 3-gram gram 1-gram gram back-off n-gram 190

6 back-off n-gram n-gram / n-gram 2000; 2001a 2.3 形態素解析の ( 古典語解析上の ) 問題点 n-gram n-gram 粒度 : + : UniDic : + : UniDic UniDic

7 : : + : : + : + + : + : 掛詞一貫性 UniDic + + UniDic : + : + : : + : 複合名詞動詞 1953;

8 表記の揺れ 3. 計量分析指標文字上の分析指標漢字率文字の相違率 ( 編集距離による計量 ) Levenshtein 6 (1993) 193

9 substitution insertion deletion ,,, 2 sentence 2 1 sentence 1 case substitute case insert delete 文体の分析指標 Jaro-Winkler distance (Miyake 2013) (Cohen, Ravikumar, and Fienberg 2003) 10 1 A B C 2 A B C 194

10 名詞の比率 Modifier Verb Ratio (MVR) MVR 3 MVR Modifier Verb Ratio 指示詞の比率 ;

11 文の長さ引用文の比率心情表現の比率各種品詞の比率語種の比率 n-gram 分析の類似性 ( パープレキシティーの利用 ) 2.2 n-gram Uzuner and

12 Katz gram 8 PP n-gram SRILM 16 8 P w 1,,w N PP n-gram 2 4 n-gram case 1 1/3 = =PP 3 case 2 1/2 =0.5 2 case 1 case 2 n-gram case 2 case 3 case 2 3 3/4 4/3 = case 頻度統計の分析指標 1970;

13 2004 Bag-of-words n-gram n-gram Bag-of-words コサイン類似度 Bag of words h i = w 1,w 2,,w N i 1 i I w n N I 0 i j c 9 T Euclid 助動詞出現頻度相関和泉式部日記 4 異本間の関係性と更級日記との比較 4.1 底本について和泉式部日記あるいは和泉式部物語 ; 2003; p

14 , d{} d{[ } d{[ ] [ } d{[ } d{[ ]@ } d{[ ; 更級日記形態素解析 UniDic 2010; 2011; 2011 MeCab

15 % 和泉式部日記の系統論 , B 和泉式部日記の文体上の特徴 :

16 結果と考察文体の分析指標 %

17 MVR (PC1) 2(PC2) 2 26 R( % 202

18 Euclid 編集距離による分析 n-gram の分析 2001b 1-gram 3-gram 3-gram

19 8 3-gram 6: 頻度統計の分析 : [%] UniDic

20 : Bag of words 11: [%] 205

21 5. まとめ

22 謝辞文献 Brinegar, C. (1963) Mark Twain and the Quintus Curtius Snodgrass Letters: A Statistical Test of Authorship, Journal of the American Statistical Association, 58: Cohen, W. W., Ravikumar, P., and Fienberg, S. E. (2003) A Comparison of String Distance Metrics for Name-Matching Tasks, in Proceedings IJCAI-03 Workshop on Information Integration, Grimmer, J. and Stewart, B. M. (2013) Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts, Political Analysis, Mingzhe, J. and Minghu, J. (2012) Text Clustering on Authorship Attribution Based on the Features of Punctuations Usage, in Proceedings International Conference on Signal Processing, 3: Miyake, M. (2013) Different Characteristics of Variant Readings Based on Comparison of Major Textual Similarity Measures, in Proceedings Japanese Association for Digital Humanities (JADH). Uzuner, O. and Katz, B. (2005) A Comparative Study of Language Models for Book and Author Recognition, in Proceedings International Joint Conference on Natural Language Processing (IJCNLP). (1944) ( ) (1991) ( ) (1956), 4(4) (1978), 10: (1981) (2002) (1957) (1956), 24: (1961) (1991) (2010) ( ), CH-85: 1 8 (2011) ( ), CH-92: 1 4 (2011) UniDic 207

23 17, (1958), 27(4) (2007), (1993) (1953), 15: (1961), 30(11) (1991) 1 (1967) (I), 1 38 (1953), 2: (1999) (2000), 48: (2007), 55: (1953), (2004) Conditional Random Fields ( ), (2005) (2000) n :, 29: (2003) (2001), 1 6 (2001a) N-gram, 2: (2001b) N-gram 7, (2011), 253: ( ) (1981) (1957) 20 (1958), 32: (1963), 53: (1986) (2001) N-gram 208

24 , 2: (1925) (2012) bigram ( 2012), (1995), 5: (1996) (2004) :, J87-D- II: (2010) (1970), 53: 1 8 (1942) (2004)? (2007), 55: (1977) (2007) CHISEConference 2005 & CodeFest 2005, 1 7 (2011), (1981) (1964)

25 Mathematical Linguistics, Vol.29 No.6 (September 2014) pp Paper Metric Japanese Study on Diary Literature in the Heian Period and Objective Measurement on Relationship between Variants Analyses on Izumishikibu nikki and Sarashina nikki Yuuki Tachioka (Mitsubishi Electric Corporation) Abstract: Stylometrics analyzes the style of texts based on some metric features. These methods have been mainly applied to modern Japanese texts, and shown its effectiveness especially for authorship attribution. However, when these methods are applied to classical literature texts, existence of variants for the same work causes problems because there are many variants for them, which rarely have an original text, and sometimes these variants are greatly different from the original one. This paper validates a method that represents a relationship between variants quantitatively, using edit distance or perplexity. Experiments on Izumishikibu nikki, which is one of the most popular diary works in the Heian period, shows that the proposed method has a better correspondence to the results shown in the previous bibliographical studies, compared to the conventional principal component analysis using multiple metric features. Furthermore, comparison with Sarashina nikki, which is another diary work in the Heian period, confirms that the difference between variants for the same work is much smaller than that between different works. Keywords: Variants, Stylometrics, Edit distance, N-gram model, Perplexity 210

pp DC 2,

pp DC 2, 計量国語学アーカイブ ID KK300501 種別論文 A タイトル近代二字漢語における同語異表記の実態と変化形態論情報付きコーパスを用いて Title Diachronic Variation in Orthography of Two-Character Sino-Japanese Words in Modern Japanese: A Corpus-based Study 著者間淵洋子

論文 pp 要旨 キーワード N 1. はじめに ; 2000; 2004 authorship attribution Brinegar 1963; 2004; Uzuner and Katz 2005; Mingzhe and Ming

論文 pp 要旨キーワード N 1. はじめに ; 2000; 2004 authorship attribution Brinegar 1963; 2004; Uzuner and Katz 2005; Mingzhe and Ming