Down Sendromuna Yönelik Türkçe X İletilerinin Yapay Zekâya Dayalı Duygu Çözümlemesi ve Niteliksel Olarak İncelenmesi

Öz

Amaç: Bu çalışmada Down sendromuna yönelik Türkçe X iletilerinin yapay zekâya dayalı duygu çözümlemesi ve niteliksel olarak incelenmesi amaçlanmıştır. Yöntem: Çalışmanın tasarımı karma yöntemdir. Ölçüt örnekleme tekniği kullanılan araştırmanın örneklemini “Down sendromu” anahtar kelimesiyle taranan 73.840 gönderi oluşturmuştur. Veriler Tweepy Paketi kullanılarak elde edilmiş olup BERT yöntemi kullanılarak özellik çıkarımı gerçekleştirilmiştir. Duygu analizi yapay sinir ağlarıyla sınıflandırılmıştır. En fazla beğeni ve paylaşım alan 738 gönderi Colaizzi’nin fenomenolojik analiz basamaklarına göre analiz edilmiştir. Bulgular: Down sendromu ile ilgili paylaşılan Türkçe iletilerin %30.23’ünün olumsuz, %39.87’sinin nötr ve %29.90’ının olumlu duygular içerdiği belirlenmiştir. Down sendromuna yönelik Türkçe X iletilerinden elde edilen bulgular “Sosyal Damgalama”, “Farkındalık ve Destek”, “Medyada Temsiliyet” ve “Sosyal Medyada Destekleyici ve Dayanışmacı Paylaşımlar” olmak üzere dört tema altında toplanmıştır. Sonuç: Bu çalışmada yapay zekâ temelli duygu çözümlemesi ile incelenen Down sendromuyla ilgili Türkçe iletilerin yaklaşık üçte birinin olumsuz duygular içerdiği saptanmıştır. Ayrıca, Down sendromu, genetik bir bozukluk yerine hastalık olarak algılanmakta ve bireyler "mutlu" ya da "melek" gibi kalıplaşmış kimliklerle tanımlanmaktadır. "Down sendromlu" ifadesi hakaret amaçlı kullanılmakta, bu bireyler ayrımcılığa uğramaktadır. Medyada gündelik yaşama katılımları olağanüstü gösterilirken, sosyal medyada farkındalık ve destek içeriği bulunmakta, aileler deneyim paylaşmaktadır. Sivil toplum kuruluşları Down sendromunu genetik bir farklılık olarak vurgulamakta, başarı öyküleri toplumsal katılımı desteklemektedir. Ayrıca, sosyal medyada oluşturulan dayanışma ağı hak taleplerini görünür kılmaktadır.

Anahtar Kelimeler

• Down sendromu
• doğal dil işleme
• makine öğrenmesi
• sosyal medya
• sosyal damgalama

Artificial Intelligence-Based Sentiment Analysis and Qualitative Analysis of Turkish X Posts about Down syndrome

Öz

Objective: This study aims to analyze Turkish X messages about Down syndrome with AI-based sentiment analysis and qualitative analysis. Method: The design of the study was mixed method. Using the criterion sampling technique, the sample of the study consists of 73,840 posts searched with the hashtag “Down syndrome”. Data extracted using Tweepy Package and feature extraction was performed using BERT method. Sentiment analysis was classified with artificial neural networks. The 738 posts with the most likes and shares analyzed according to Colaizzi's phenomenological analysis steps. Results: Among the Turkish messages about Down syndrome, 30.23% were negative, 39.87% were neutral and 29.90% were positive. The findings obtained from Turkish X messages about Down syndrome were grouped under four themes: ‘Social Stigmatisation’, ‘Awareness and Support’, ‘Representation in Media’ and ‘Supportive and Solidarity-Oriented Posts on Social Media’. Conclusion: Approximately one-third of the Turkish messages about Down syndrome analyzed with artificial intelligence-based sentiment analysis contain negative sentiments. In addition, Down syndrome perceived as a disease rather than a genetic disorder and individuals identified with stereotypical identities such as “happy” or “angel”. The term “Down syndrome” continues to be used as an insult and individuals discriminate against people with Down syndrome. While their participation in daily life portrayed as extraordinary in the media, awareness and support content appears on social media and families share stories about their experiences. Non-governmental organizations emphasize that Down syndrome is a genetic difference and success stories support social participation. In addition, the social media solidarity network makes the demands for rights visible.

Anahtar Kelimeler

• Down syndrome
• natural language processing
• machine learning
• social media
• social stigmatization

Kaynakça

1. Alhaddad MH, Anwer F, Basonbul RA, Butt NS, Noor MI, Malik AA (2018) Knowledge and attitude towards Down syndrome among people in Jeddah, Saudi Arabia. Proc Shaikh Zayed Postgrad Med Inst, 32:56-65.
2. Antonarakis SE, Lyle R, Dermitzakis ET, Reymond A, Deutsch S (2020) Chromosome 21 and down syndrome: From genomics to pathophysiology. Nat Rev Genet, 11:885-899.
3. Budenz A, Klassen A, Purtle J, Yom Tov E, Yudell M, Massey P (2020) Mental illness and bipolar disorder on twitter: Implications for stigma and social support. J Ment Health, 29:191–199.
4. Burch L (2017) A world without down’s syndrome? Online resistance on twitter: #worldwithoutdowns and #justaboutcoping. Disabil Soc, 32:1085-1089.
5. Cho H, Kim KM, Kim JY, Youn BY (2024) Twitter discussions on #digitaldementia: Content and sentiment analysis. J Med Internet Res, 26:1-17.
6. Colaizzi PF (1978) Psychological research as the phenomenologist views it. In: Existential-Phenomenological Alternatives for Psychology, (Eds RS Valle, King M).:48–71. New York, Oxford University Press.
7. Creswell JW (2020) Felsefi varsayımlar ve yorumlayıcı çatılar [Philosophical assumptions and interpretive frameworks]. In Nitel araştırma yöntemleri: Beş yaklaşıma göre nitel araştırma ve araştırma deseni, 2nd ed. (Eds Bütün M, Demir SB):15–42. Ankara, Siyasal Publisher.
8. Creswell JW, Plano Clark VL (2017) Designing and Conducting Mixed Methods Research, 3rd ed. Thousand Oaks, Sage.

9. de Graaf G, Buckley F, Skotko BG (2021) Estimation of the number of people with down syndrome in the United States. Genet Med, 23:629-635.
10. Dikeç G, Oban V, Usta MB (2023) Şizofreniye yönelik Türkçe twitter iletilerinin nitel ve yapay zekâ temelli duygu çözümlemesi. Turk Psikiyatri Derg, 34:145-153.
11. Doğan MB, Oban V, Dikeç G (2023) Qualitative and artificial ıntelligence-based sentiment analyses of anti-LGBTI+ hate speech on twitter in Turkey. Issues Ment Health Nurs, 44:112–120.
12. Edgerton R (1980) The study of deviance: marginal man or everyman? In The Making of Psychological Anthropology, (Ed Spindler GD):444–476. Berkeley, University of California Press.
13. Evans HH, Rice ST (2018) Angel unaware and down syndrome awareness. Pediatrics, 142:1-5.
14. Feldman R (2013) Techniques and applications for sentiment analysis. Commun ACM, 56:82-89.
15. Fetters MD, Curry LA, Creswell JW (2013) Achieving integration in mixed methods designs: Principles and practices. Health Serv Res, 48:2134-2156.
16. Franzke AS, Muis I, Schäfer MT (2021) Data Ethics Decision Aid (DEDA): A dialogical framework for ethical inquiry of AI and data projects in the Netherlands. Ethics Inf Technol, 22:1-17.
17. Gilmore L, Campbell J, Cuskelly M (2003) Developmental expectations, personality stereotypes, and attitudes towards inclusive education: Community and teacher views of Down syndrome. Int J Disabil Dev Educ, 50:65-76.
18. Goggin G, Newell C (2020) Disability in Australia: Exposing Social Apartheid, 1st ed. London, Routledge.
19. Göksel P, Oban V, Dikeç G, Usta MB (2023) Qualitative and artificial intelligence-based sentiment analysis of Turkish Twitter messages related to Autism Spectrum Disorders. Cureus, 15:1-7.
20. Heinemann W (1990) Meeting the handicapped: A case of affective-cognitive inconsistency. Eur Rev Soc Psychol, 1:323-338.
21. Husserl E (1960) Cartesian Meditations: An Introduction to Phenomenology (Trans Cairns D). Dordrecht, Springer.
22. Johnson B, West R (2021) Ableist contours of down syndrome in Australia: Facebook attitudes towards existence and parenting of people with down syndrome. J Sociol, 57:286-304.
23. Kurtgöz A, Genç M (2024) Spiritual care perspectives of elderly ındividuals with parkinson's disease and formal caregivers: A qualitative study in Turkish nursing homes. J Relig Health, 63:2106–2124.
24. Lincoln YS, Guba EG (1985) Naturalistic Inquiry, 1st ed. Newbury Park, Sage.
25. Pace JE, Shin M, Rasmussen SA (2010) Understanding attitudes toward people with Down syndrome. Am J Med Genet A, 152:2185-2192.
26. Pary RJ (2004) Behavioral and psychiatric disorders in children and adolescents with Down syndrome. Ment Health Asp Dev Disabil, 7:69-76.
27. Pavlova A, Berkers P (2022) "Mental Health" as defined by Twitter: Frames, emotions, stigma. Health Commun, 37:637–647.
28. Polit DF, Beck CT (2006) Essentials of Nursing Research, 6th ed. Philadelphia, Lippincott Williams & Wilkins.
29. Rezaee K (2024) You look at the face of an angel: An ınnovative hybrid deep learning approach for detecting Down Syndrome in children's faces through facial analysis. J AI Data Min, 12:287-303.
30. Rodríguez Díaz B (2019) Mission impossible? Preventing discrimination on grounds of disability of foetuses with Down syndrome in Spain after the emergence of non-invasive prenatal testing. Int J Discrim Law, 19:178-199.
31. Rodríguez ND, Mateo EA, Rodríguez VB, Rodríguez-Pérez A (2018) Intergroup trust and anxiety: The two sides of stigma towards people with Down syndrome. An Psicol, 34:117-122.
32. Roesslein J (2009) Tweepy documentation. http://docs.tweepy.org/en/v3.5.0/ (Accessed 01.11.2024).
33. Rogers K, Ziviani J (2022) Social media representations of Down syndrome: A content analysis of Twitter discussions. J Intellect Disabil Res, 66:421-435.
34. Rooney NM (2014) Promoting positive attitudes toward individuals with Down syndrome: The relationship between indirect contact interventions and the quality of previous contact (Honors thesis). Minnesota, Macalester College.
35. Sani-Bozkurt S (2018) Identifying network structure, influencers and social mood in digital spheres: A sentiment and content analysis of Down syndrome awareness. World J Educ Technol Curr Issues, 10:10-19.
36. Schweter S (2020) BERTurk - BERT models for Turkish (Version 1.0.0). Zenodo. Retrieved from https://zenodo.org/record/3770924 (Accessed 30.05.2024)
37. Search Engine Land (2021) BERT now used on almost every English query. https://searchengineland.com/google-bert-used-on-almost-every-englishquery-342193 (Accessed 01.11.2024)
38. Taylor DC (2013) Stigma and prejudice in the language of sickness. Epilepsy Behav, 27:204-205.
39. Tong A, Sainsbury P, Craig J (2007) Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care, 19:349–357.
40. Tremain S (2022) Foucault and the Government of Disability, 2nd ed. Ann Arbor, University of Michigan Press.
41. Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, Poole C, Schlesselman JJ, Egger M, STROBE Initiative (2014) Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. Int J Surg, 12:1500–1524.
42. Werner S (2015) Public stigma in intellectual disability: do direct versus indirect questions make a difference? J Intellect Disabil Res, 59:958–969.
43. Zerres K, Rudnik-Schöneborn S, Holzgreve W (2021) Do non-invasive prenatal tests promote discrimination against people with Down syndrome? What should be done? J Perinat Med, 49:965-971.