The Korean Association for the Study of English Language and Linguistics

[ Article ]

Korea Journal of English Language and Linguistics - Vol. 23, No. 0, pp.431-447

ISSN: 1598-1398 (Print) 2586-7474 (Online)

Print publication date 30 Jan 2023

Received 03 Apr 2023 Revised 17 May 2023 Accepted 07 Jun 2023

DOI: https://doi.org/10.15738/kjell.23..202306.431

Detecting Suicide Notes with the Probability of Positive Sentiment and Interquartile Range

Yong-hun Lee

Instructor, Dept. of Linguistics, Chungnam Naional University, Tel: 042) 821-6391 yleeuiuc@hanmail.net

© 2023 KASELL All rights reserved
This is an open-access article distributed under the terms of the Creative Commons License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

This paper proposes a new algorithm for detecting suicide notes using sentiment analysis. As suicides increase nowadays, it is important to detect the suicide signs before the actual suicides are committed. Detecting suicide signs is not so easy, because suicide notes are usually short. This study proposes a modified algorithm of sentiment analysis which is based on the probability of positive sentiments (PPS), not on the categorical classifications. The original BERT model is revised so that the model calculates the PPS values for each sentence. A total of 8 corpora are constructed, among which 4 are the corpora of suicide notes, and the others are novels. For each sentence in the corpora, the PPS values are calculated using the revised BERT model. Then, the distributions of PPSs are statistically analyzed with Interquartile Range (IQR). The suicide notes are distinguished with more than 30 IQR values. In the experiments with the corpora of suicide notes and ordinary texts, the developed method achieves about 86% of accuracy. The proposed algorithm can make use of the sentiment properties of suicide notes, and it is effective not only for large-size corpora but also for small-size suicide notes.

Keywords:

suicide notes, sentiment analysis, BERT, probability of positive sentiment, IQR

References

Chaski, C. 2012. Author identification in the forensic setting. In L. Solan and P. Tiermsa, eds., The Oxford Handbook of Forensic Linguistics, 333-372. Oxford, UK: Oxford University Press. [https://doi.org/10.1093/oxfordhb/9780199572120.013.0036]
Coulthard, M. and A. Johnson. 2016. An Introduction to Forensic Linguistics. Cambridge, MA: Cambridge University Press.
Desmet, M. and V. Hoste. 2013. Emotion detection in suicide notes. Expert Systems with Applications 40(16), 6351-6358. [https://doi.org/10.1016/j.eswa.2013.05.050]
Devlin, J., M. Chang, K. Lee. and K. Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. arXiv Preprint arXiv:1810.04805, .
Durkheim, E. 1951. Suicide. New York: The Free Press.
Edelman, A. and L. Renshaw. 1982. Genuine versus simulated suicide notes: An issue revisited through discourse analysis. Suicide and Life-Threatening Behavior 12(2), 103-113. [https://doi.org/10.1111/j.1943-278X.1982.tb00917.x]
Ghosh, S., A. Ekbal, and P. Bhattacharyya. 2020. CEASE: A corpus of emotion-annotated suicide notes in English. In Proceedings of the 12th Language Resources and Evaluation Conference, 1618-1626. Marseille, France, 11-16 May 2020.
Giles, S. 2007. The Final Farewell: Using a Narrative Approach to Explore Suicide Notes as Ultra-social Phenomenon. Unpublished doctoral dissertation, University of Liverpool, Liverpool, England.
Lee, Y. 2021. English Island Constraints Revisited: Experimental vs. Deep Learning Approach. English Language and Linguistics 27(3), 21-45.
Lee, Y. 2022. Negative Polarity Items in English: A Deep Learning Model and Statistical Analysis. Korean Journal of Linguistics 47(1), 29-56.
Lee, Y. and G. Joh. 2019. Identifying suicide notes using forensic linguistics and machine learning. The Linguistic Association of Korean Journal 27(2), 171-191. [https://doi.org/10.24303/lakdoi.2019.27.2.171]
Leenaars, A. 1988. Suicide Notes. New York: Human Sciences Press.
Matykiewicz, P., D. Wlodzislaw. and J. Pestian. 2009. Clustering semantic spaces of suicide notes and newsgroups articles. In Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing, 179-184. Boulder, Colorado, USA, 5 June 2009. [https://doi.org/10.3115/1572364.1572389]
McCart, J., D. Finch, J. Jarman, E. Hickling, J. Lind, M. Richardson, D. Berndt. and S. Luther. 2012. Using ensemble models to classify the sentiment expressed in suicide notes. Biomedical Informatics Insights 5(1), 77-85. [https://doi.org/10.4137/BII.S8931]
Mitchell, T. 1997. Machine Learning. New York: McGraw Hill.
Olsson, J. 2004. Forensic Linguistics: An Introduction to Language, Crime, and the Law. London, UK: Continuum.
Olsson, J. 2008. Forensic Linguistics: An Introduction to Language, Crime, and the Law, 2nd edition. London, UK: Continuum.
Pennebaker, J. and L. King. 1999. Linguistic styles: Language use as an individual difference. Journal of Personality and Social Psychology 1999 77(6), 1296-1312. [https://doi.org/10.1037/0022-3514.77.6.1296]
Pennebaker, W., E. Francis. and J. Booth. 2001. Linguistic Inquiry and Word Count (LIWC): LIWC2001. Mahwah, NJ: Lawrence Erlbaum Associates.
Pestian, J., P. Matykiewicz, J. Grupp-Phelan, A. Lavanier. J. Combs. and R. Kowatch. 2008. Using natural language processing to classify suicide notes. In Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing (BioNLP’08), 96-99. Columbus, Ohio, USA, 19 June 2008. [https://doi.org/10.3115/1572306.1572327]
Pestian, J., P. Matykiewicz. and M. Linn-Gust. 2012a. What’s in a note: Construction of a suicide note corpus. Biomedical Informatics Insights 5(5), 1-6. [https://doi.org/10.4137/BII.S9042]
Pestian, J., P. Matykiewicz, M. Linn-Gust, B. South, O. Uzuner, J. Wiebe, B. Cohen, J. Hurdle. and C. Brew. 2012b. Sentiment analysis of suicide notes: A shared task. Biomedical Informatics Insights 5(1), 3-16. [https://doi.org/10.4137/BII.S9042]
Pestian, J., H. Nasrallah, P. Matykiewicz, A. Bennett. and A. Leenaars. 2010. Suicide note classification using natural language processing: A content analysis. Biomedical Informatics Insights 3(3), 19-28. [https://doi.org/10.4137/BII.S4706]
Roubidoux, S. 2012. Linguistic Manifestations of Power in Suicide Notes: An Investigation of Personal Pronouns. Unpublished doctoral dissertation, University of Wisconsin at Oshkosh, Oshkosh, Wisconsin, USA.
Samuel, A. 1959. Some Studies in Machine Learning Using the Game of Checkers. IBM Journal 3, 210-229. [https://doi.org/10.1147/rd.33.0210]
Sboev, A., D. Gudovskikh, R. Rybka, and I. Moloshnikov. 2015. A quantitative method of text emotiveness evaluation on base of the psycholinguistic markers founded on morphological features. Procedia Computer Science 66, 307-316. [https://doi.org/10.1016/j.procs.2015.11.036]
Shapero, J. 2011. The Language of Suicide Notes. Unpublished doctoral dissertation, University of Birmingham, Birmingham, UK.
Sheidman, E. and N. Faberow. 1963. Clues to Suicide. New York: McGraw-Hill.
Shneidman, S. 1996. The Suicidal Mind. Oxford, UK: Oxford University Press.
Svartvik, J. 1968. The Evans Statements: A Case for Forensic Linguistics. Gothenburg, Sweden: University of Gothenburg Press.
Tausczik, Y. and J. Pennebaker. 2010. The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology 29(1), 24-54. [https://doi.org/10.1177/0261927X09351676]
Vaswani, A., N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. Gomez, L. Kaiser. and I. Polosukhin. 2017. Attention is all you need. arXiv Preprint arXiv:1706.03762, .
Wang, W., L. Chen, M. Tan, S. Wang. and A. Sheth. 2012. Discovering fine-grained sentiment in suicide notes. Biomedical Informatics Insights 5(1), 137-145. [https://doi.org/10.4137/BII.S8963]
Yang, H., A. Willis, A. De Roeck. and B. Nuseibeh. 2012. A hybrid model for automatic emotion recognition in suicide notes. Biomedical Informatics Insights 5(1), 17-30. [https://doi.org/10.4137/BII.S8948]