Elamur Research⌕
☰
home / papers / 2000 / Automatic Text Categorization In Terms Of Genre and Author

Automatic Text Categorization In Terms Of Genre and Author

Paper · 2000 · Efstathios Stamatatos, Nikos Fakotakis and George Kokkinakis

Summary

The paper addresses the classification of texts by style rather than content, covering both genre and authorship for Modern Greek. Earlier stylometry mostly relied on lexical statistics or deliberately avoided NLP tools. The authors instead extract style markers from an existing sentence and chunk boundaries detector (SCBD). They add a new type of feature, analysis-level markers, which describe how the tool processed the text, for example how difficult the parsing was. These features come at no extra computational cost.

The 22 markers are grouped into token, syntax and analysis levels. They are fed to statistical classifiers, including multiple regression. The system is tested on genre detection, author identification and author verification, using unedited web texts with no manual cleaning or sampling.

The method beats vocabulary-richness functions and frequencies of the most common words. Genre results are noticeably better than authorship results. The approach matters because it is trainable, cheap and quick to adapt to new categories. It also shows how other existing NLP tools could be reused for stylometry.

Key points

  • The feature set has 22 markers (M01–M22) on three levels: token, syntax and analysis. All are computed automatically by the SCBD for Modern Greek.
  • In the regression models, token-level markers contribute most. The average absolute t value is 1.67 for genre and 1.88 for authorship. Analysis-level markers come next (1.11 / 1.01), and syntax-level markers contribute least (0.75 / 0.86).
  • For genre detection, the most informative individual markers are M02, M12 and M15.
  • Deliberately degrading the SCBD's boundary detection raised the genre error to about 0.23, roughly 25% worse. Accuracy therefore depends on the quality of the underlying tool.
  • About 10 training texts per category already give fairly high accuracy. Texts of at least 1,000 words help, especially for author identification, yet most texts in the corpora were shorter than that. The authors suggest combining the method with lexical approaches such as CWF-30.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of PatrasPátrai · Greece2 works · 432 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

💡 Concepts 1Key ideas and methods explained on our concept pages.Text classification · 50 works

Cite

Cite as Efstathios Stamatatos, Nikos Fakotakis and George Kokkinakis. 2000. Automatic Text Categorization In Terms Of Genre and Author. Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work.

@article{stamatatos2000automatic,
  title = {{Automatic Text Categorization In Terms Of Genre and Author}},
  author = {Efstathios Stamatatos, Nikos Fakotakis and George Kokkinakis},
  journal = {Computational Linguistics},
  year = {2000},
  doi = {10.1162/089120100750105920},
  url = {https://doi.org/10.1162/089120100750105920}
}

References

41 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Baayen, Harald, Hans Van Halteren, and Fiona Tweedie. 1996. Outside the cave of shadows: Using syntactic annotation to enhance authorship attribution. Literary and Linguistic Computing, 11(3):121-131. find
  2. Biber, Douglas. 1990. Methodological issues regarding corpus-based analyses of linguistic variations. Literary and Linguistic Computing, 5:257-269. find
  3. Biber, Douglas. 1993a. Representativeness in corpus design. Literary and Linguistic Computing, 8:1-15. find
  4. Biber, Douglas. 1993b. Using register-diversified corpora for general language studies. Computational Linguistics, 19(2):219-242. find
  5. Biber, Douglas. 1995. Dimensions of Register Variation: A Cross-Linguistic Comparison. find
  6. Brainerd, Barron. 1974. Weighting Evidence in Language and Literature: A Statistical Approach. University of Toronto Press. find
  7. Brinegar, Claude S. 1963. Mark Twain and the Quintus Curtius Snodgrass letters: A statistical test of authorship. Journal of the American Statistical Association, 58:85-96. find
  8. Brunet, Ettienne. 1978. Vocabulaire de Jean Giraudoux: Structure et Evolution. Slatkine. find
  9. Burrows, John F. 1987. Word-patterns and story-shapes: The statistical analysis of narrative style. Literary and Linguistic Computing, 2(2):61-70. find
  10. Burrows, John F. 1992. Not unless you ask nicely: The interpretative nexus between analysis and information. Literary and Linguistic Computing, 7(2):91-109. find
  11. Edwards, Allen F. 1979. Multiple Regression and the Analysis of Variance and Covariance. find
  12. W. H. Freeman, San Francisco, CA. find
  13. Eisenbeis, Robert A., and Robert B. Avery. 1972. Discriminant Analysis and Classification Procedures: Theory and Applications. D.C. Health and Co., Lexington, MA. find
  14. Fakotakis, Nikos, Anastasios Tsopanoglou, and George Kokkinakis. 1993. A text-independent speaker recognition system based on vowel spotting. Speech Communication, 12:57-68. find
  15. Holmes, David I. 1992. A stylometric analysis of Mormon scripture and related texts. Journal of the Royal Statistical Society, Series A, 155(1):91-120. find
  16. Holmes, David I. 1994. Authorship attribution. Computers and the Humanities, 28:87-106. find
  17. Holmes, David I., and Richard S. Forsyth. 1995. The Federalist revisited: New directions in authorship attribution. find
  18. HonorG Antony. 1979. Some simple measures of richness of vocabulary. find
  19. Association for Literary and Linguistic Computing Bulletin, 7(2):172-177. find
  20. Karlgren, Jussi. 1999. Stylistic experiments in information retrieval. In T. Strzalkowski, editor, Natural Language Information Retrieval. Kluwer Academic Publishers, pages 147-166. find
  21. Karlgren, Jussi and Douglass Cutting. 1994. Recognizing text genres with simple metrics using discriminant analysis. In Proceedings of the 15th International Conference on Computational Linguistics (COLING '94), pages 1,071-1,075. find
  22. Kessler, Brett, Geoffrey Nunberg, and Hinrich Schiitze. 1997. Automatic detection of text genre. In Proceedings of 35th Annual Meeting, pages 32-38. Association for Computational Linguistics.
  23. Michos, Stefanos, Efstathios Stamatatos, Nikos Fakotakis, and George Kokkinakis. 1996. An empirical text categorizing computational model based on stylistic aspects. In Proceedings of the 8th Conference on Tools with Artificial Intelligence (ICTAI'96), pages 71-77. find
  24. Morton, Andrew Q. 1965. The authorship of Greek prose. Journal of the Royal Statistical Society, Series A, 128:169-233. find
  25. Mosteller, Fredrick and David Wallace. 1984. find
  26. Applied Bayesian and Classical Inference: The Case of the Federalist Papers. Addison-Wesley, Reading, MA. find
  27. Oakman, Robert L. 1980. Computer Methods for Literary Research. University of South Carolina Press, Columbia. find
  28. Sichel, Herbert S. 1975. On a distribution law for word frequencies. Journal of the American Statistical Association, 70:542-547. find
  29. Simpson, Edward H. 1949. Measurement of diversity. Nature, 163:688. find
  30. Smith, M. W. A. 1983. Recent experience and new developments of methods for the determination of authorship. find
  31. Association for Literary and Linguistic Computing Bulletin, 11:73-82. find
  32. Smith, M. W. A. 1985. An investigation of Morton's method to distinguish Elizabethan playwrights. Computers and the Humanities, 19(1):3-21. find
  33. Stamatatos, Efstathios, Nikos Fakotakis, and George Kokkinakis. 1999. Automatic extraction of rules for sentence boundary disambiguation. In Proceedings of the Workshop in Machine Learning in Human Language Technology, Advance Course on Arti~'cial Intelligence (ACAI'99), pages 88-92. find
  34. Stamatatos, Efstathios, Nikos Fakotakis, and George Kokkinakis. 2000. A practical chunker for unrestricted text. In Proceedings of the 2nd International Conference on Natural Language Processing, pages 139-150. find
  35. Strzalkowski, Tomek. 1994. Robust text processing in automated information retrieval. In Proceedings of the 4th Conference On Applied Natural Language Processing (ANLP'94), pages 168-173. find
  36. Tabachnick, Barbara G. and Linda S. Fidell. 1996. Using Multivariate Statistics. Third edition. HarperCollins College Publishers. find
  37. Tweedie, Fiona and Harald Baayen. 1998. How variable may a constant be? Measures of lexical richness in perspective. Computers and the Humanities, 32(5):323-352. find
  38. Tweedie, Fiona, Sameer Singh, and David I. Holmes. 1996. Neural network applications in stylometry: The Federalist Papers. Computers and the Humanities, 30(1):1-10. find
  39. Yang, Yiming. 1999. An evaluation of statistical approaches to text categorization. Information Retrieval Journal, 1(1):69-90. find
  40. Yang, Yiming and Xin Liu. 1999. A re-examination of text categorization methods. In Proceedings of the ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR'99), pages 42–49. find
  41. Yule, George U. 1944. The Statistical Study of Literary Vocabulary. Cambridge University Press. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/stamatatos-2000-automatic-text-categorization-terms-genre/