SAĞLIKLI VE KARACİĞER HASTALIĞI OLAN BİREYLERİN DOĞRUSAL REGRESYON SINIFLANDIRMA ALGORİTMASIYLA TAHMİN EDİLMESİ

Year 2023, Volume: 6 Issue: 3, 229 - 233, 24.10.2023

Handan Tanyıldızı Kökkülünk

https://doi.org/10.26650/JARHS2023-1231512

Abstract

Amaç: Bu çalışmada sağlıklı ve karaciğer hastalığı olan bireylere ait bazı kan biyokimya parametreleri (ALB, ALP, ALT, AST, BIL, CHE, CHOL, CREA, GGT ve PROT), cinsiyet ve yaş bilgileri kullanılarak karaciğer hastalığı yok/ var kategorik tahmini yapılması amaçlanmıştır.
Gereç ve Yöntem: R Studio programında makine öğrenmesine ait çoklu doğrusal regresyon ile tahmin elde edilmiştir. Akaike bilgi kriteri kullanılarak tahmin üzerine yüksek katkısı olan parametreler seçilerek makine öğrenmesinde iyileştirilmeye gidilmiştir.
Bulgular: Tahmine pozitif yönlü etkisi olan en güçlü 3 parametre sırasıyla AST, BIL ve GGT; negatif yönlü etkisi olan en güçlü 3 parametre sırasıyla CHOL, CHE ve ALB bulunmuştur. Kullanılan modelin doğruluğu %91, kesinlik %99, geri çağırma 0,91 ve F skoru %94 olarak bulunmuştur. Korelasyon ilişkisi grafiği incelendiğinde AST ‘nin sağlıklı/karaciğer hastası bireylerde güçlü bir ayırıcı parametre olduğu tespit edilmiştir.
Sonuç: Çoklu doğrusal regresyonun, kategorik hastalık sınıflandırması için tercih edilebilir bir yöntem olduğu bulunmuştur.

Keywords

Makine öğrenmesi, karaciğer, sınıflandırma

ESTIMATION OF HEALTHY AND LIVER DISEASED INDIVIDUALS BY A LINEAR REGRESSION CLASSIFICATION ALGORITHM

Abstract

Objective: In this study, the aim was to make a categorical estimation of the absent/presence of liver disease by using some blood biochemistry parameters (ALB, ALP, ALT, AST, BIL, CHE, CHOL, CREA, GGT, and PROT), gender and the age of healthy individuals, and those with liver disease.
Material and methods: The prediction was obtained with multiple linear regression of machine learning in the R Studio program. Machine learning was improved by selecting parameters that have a high contribution to the prediction by using the Akaike information criterion.
Results: The three strongest parameters with a positive effect on the estimation were AST, BIL, and GGT, respectively; The three strongest parameters with negative effects were CHOL, CHE, and ALB, respectively. The accuracy of the model used was 91%, the precision was 99%, the recall was 0.91, and the F score was 94%. When the correlation relationship graph was examined, it was determined that AST was a strong differential parameter in healthy/liver diseased individuals.
Conclusion: Multiple linear regression is a preferable method for categorical disease classification.

Keywords

Machine learning, liver, classification

References

• Magoulas GD, Prentza A. Machine learning in medical applications. In: Paliouras G, Karkaletsis V, Spyropoulos CD, editors. Machine learning and its applications: advanced lectures. Berlin, Heidelberg: Springer; 2001 (cited 2022) p.300–7. (Lecture Notes in Computer Science). Available from:https://doi.org/10.1007/3-540-44673-7_19. google scholar
• Stausberg J, Person M. A process model of diagnostic reasoning in medicine. Int J Med Inform 1999;54(1):9-23. google scholar
• B. Zupan, J. Halter, M. Bohanec. Qualitative model approach to computer assisted reasoning in physiology. Computer Science 1998 (cited 2022 September 2) https://www.semanticscholar. org/paper/Qualitative-Model-Approach-to-Computer-Assistedin- Zupan-Halter/4197bc7fc5af6754e99d39c204eef80a99e324c3 google scholar
• Gindi GR, Darken CJ, O’Brien KM, Stetz ML, Deckelbaum LI. Neural network and conventional classifiers for fluorescence-guided laser angioplasty. IEEE Trans Biomed Eng 1991;38(3):246-52. google scholar
• Srinivas S. A machine learning-based approach for predicting patient punctuality in ambulatory care centers. Int J Environ Health Res 2020;17(10):3703. google scholar
• Anusuya V, Gomathi V. An efficient technique for disease prediction by using enhanced machine learning algorithms for categorical medical dataset. I Inf Technol Control 2021;50(1) :102-22. google scholar
• Uddin S, Khan A, Hossain ME, Moni MA. Comparing different supervised machine learning algorithms for disease prediction. BMC Medical Inform Decis Mak 2019;19(1):281. google scholar
• Petrie JR, Guzik TJ, Touyz RM. Diabetes, hypertension, and cardiovascular disease: clinical insights and vascular mechanisms. Can J Cardiol 2018;34(5):575-84. google scholar
• Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin 2005;55(2):74-108. google scholar
• Jayaswal ANA, Levick C, Selvaraj EA, Dennis A, Booth JC, Collier J, et al. Prognostic value of multiparametric magnetic resonance imaging, transient elastography and blood-based fibrosis markers in patients with chronic liver disease. Liver Int 2020;40(12):3071-82. google scholar
• Abebe M, Melku M, Enawgaw B, Birhan W, Deressa T, Terefe B, et al. Reference intervals of routine clinical chemistry parameters among apparently healthy young adults in Amhara National Regional State, Ethiopia. Plos one 2018;13(8):e0201782. google scholar
• Hepatitis c prediction dataset. 2021 (cited 2022 August 1): 1(1):(1 screen). Available from:URL: https://www.kaggle.com/datasets/ fedesoriano/hepatitis-c-dataset. google scholar
• Yee MM, Aung EE, Khaing YM. Forecasting stock market using multiple linear regression. IJTSRD 2019;3(5):2174-6. google scholar
• Giacomino A, Abollino O, Malandrino M, Mentasti E. The role of chemometrics in single and sequential extraction assays: a review. Part II. Cluster analysis, multiple linear regression, mixture resolution, experimental design and other techniques. Anal Chim Acta 2011;688(2):122-39. google scholar
• Khalid A, Sarwat AI. Unified univariate-neural network models for lithium-ion battery state-of-charge forecasting using minimized akaike information criterion algorithm. IEEE Access 2021;9:39154-70. google scholar
• Hasan M, Islam MdM, Zarif MII, Hashem MMA. Attack and anomaly detection in IoT sensors in IoT sites using machine learning approaches. Internet of Things 2019;7:100059. google scholar
• Reference ranges for blood tests. In: Wikipedia. 2022 (cited 2022 September 1). Available from: URL:https://en.wikipedia. org /w/index.php?title=Reference_ranges_for_blood_ tests&oldid=1109845763. google scholar
• Teke M. Prediction of liver diseases with machine learning method. SMUTGD 2022;5(1):115-22. google scholar
• Akter S, Shekhar HU, Akhteruzzaman S. Application of biochemical tests and machine learning techniques to diagnose and evaluate liver disease. Adv Biosci Biotechnol 2021;12(6):154-72. google scholar
• Rahman AKM, Shamrat FM, Tasnim Z, Roy J, Hossain S. A comparative study on liver disease prediction using supervised machine learning algorithms. Int J Sci Technol Res 2019;8(11):419-22. google scholar
• Schiff ER, Maddrey WC, Reddy KR. Schiff’s Diseases of the Liver. 12th Edition. USA: Wiley-Blackwell; 2017. pp.135-218. google scholar
• Sorich MJ, Miners JO, McKinnon RA, Winkler DA, Burden FR, Smith PA. Comparison of linear and nonlinear classification algorithms for the prediction of drug and chemical metabolism by human udp-glucuronosyltransferase isoforms. J Chem Inf Comput Sci 2003;43(6):2019-24. google scholar
• Saygın E, Baykara M. Karaciğer yetmezliği teşhisinde özellik seçimi kullanarak makine öğrenmesi yöntemlerinin başarılarının ölçülmesi. FÜMBD 2021;33(2):367-77. google scholar