Application of Ensemble Learning Based Multivariate Pattern Analysis On Clinical Decision Support Systems

Abstract

Multivariable pattern analysis (MVPA) is a powerful technique that is used widely to examine clinical data. Machine learning methods are generally used in MVPA applications. Ensemble learning algorithms increase the classification performance by combining a large number of machine learning methods. Cervical cancer that is the second most common cancer type in women is a major cause of death in low-income countries. Cervical cancer, which has no symptoms in its initial stages, can be completely cured if it is diagnosed early. In this study a MVPA application is performed that uses random forests, extremely randomized trees and Adaboost ensemble learning methods to estimate the risk of cervical cancer of the patients by using the various risk factors for different screening strategies. Prediction performance of the extremely randomized trees for the cervical cancer risk was 99%, 96%, 97% and 97% for the Hinselmann, Schiller, citology and biopy screening methods.

Keywords

• Ensemble learning algorithms
• Random forests
• Extremely randomized trees
• Adaboost
• Clinical decision support systems

Kolektif Öğrenmeye Dayalı Çok Değişkenli Desen Analizinin Klinik Karar Destek Sistemlerinde Uygulanması

Abstract

Çok değişkenli desen analizi (ÇDDA) klinik verilerin incelenmesi için yaygın olarak kullanılan güçlü bir tekniktir. ÇDDA uygulamalarında genellikle makine öğrenmesi yöntemleri kullanılmaktadır. Kolektif öğrenme algoritmaları çok sayıda makine öğrenmesi metodunu bir araya getirilerek sınıflama performansını arttırmaktadır. Kadınlarda en sık görülen ikinci kanser türü olan serviks kanseri düşük gelirli ülkelerde önemli bir ölüm nedenidir. İlk evrelerinde belirti göstermeyen serviks kanseri erken teşhis edildiğinde tamamen tedavi edilebilmektedir. Bu çalışmada rastgele ormanlar, aşırı rassal ağaçlar ve Adaboost kolektif öğrenme yöntemleri kullanılarak hastaların serviks kanseri riskini çeşitli risk faktörlerinden faydalanarak farklı tarama yöntemleri açısından tahmin eden bir ÇDDA uygulaması yapılmıştır. Aşırı rassal ağaçlar algoritmasından Hinselmann, Schiller, sitoloji ve biyopsi tarama yöntemlerinin hastaların serviks kanseri olması riskini tahmin başarısı sırasıyla %99, %96, %97 ve %97 olmuştur.

Keywords

• Kolektif öğrenme algoritmaları
• Rastgele ormanlar
• Aşırı rassal ağaçlar
• Adaboost
• Klinik karar destek sistemleri

References

1. [1] M.A. Musen, B. Middleton, R.A. Greenes, “Clinical Decision-Support Systems,” Biomedical Informatics, 4. baskı. London: Springer, 2014, böl. 20, ss. 643-674.
2. [2] J.M. Hardin ve D.C. Chhieng, “Data Mining and Clinical Decision Support Systems,” Clinical decision support systems, 2. baskı. New York: Springer Science+ Business Media, LLC, 2007, böl. 3, ss. 44-63.
3. [3] N. Ye, C. Liu, P. Shi, “Metabolomics Analysis of Cervical Cancer, Cervical Intraepithelial Neoplasia and Chronic Cervicitis by 1H NMR Spectroscopy,” European Journal of Gynaecological Oncology, c. 36, s. 2, ss. 174-180, 2015.
4. [4] K. Fernandes, J.S. Cardoso, J. Fernandes, “Transfer Learning with Partial Observability Applied to Cervical Cancer Screening,” 8th Iberian Conference on Pattern Recognition and Image Analysis, Faro, Portugal, 2017, ss. 243-250.
5. [5] A. Bandyopadhyay, U. Mukherjee, S. Ghosh, S. Ghosh, S.K. Sarkar, “Pattern of Failure with Locally Advanced Cervical Cancer– A Retrospective Audit and Analysis of Contributory Factors,” Asian Pacific Journal of Cancer Prevention, c. 19, s. 1, ss. 73-79, 2017.
6. [6] W.A. Leyden, M.M Manos, A.M. Geiger, S. Weinmann, J. Mouchawar, K. Bischoff, M.U. Yood, J. Gilbert, S.H. Taplin, “Cervical Cancer in Women with Comprehensive Health Care Access: Attributable Factors in the Screening Process,” Journal of the National Cancer Institute, c. 97, s. 9, ss. 675–683, 2005.
7. [7] P. Bountris, M. Haritou, A. Pouliakis, N. Margari, M. Kyrgiou, A. Spathis, A. Pappas, I. Panayiotides, E.A. Paraskevaidis, P. Karakitsos, D.D. Koutsouris, “An intelligent clinical decision support system for patient-specific predictions to improve cervical intraepithelial neoplasia detection,” BioMed Research International, c. 2014, Article Number: 341483, 2014.
8. [8] A. Demirhan, “Neuroimage‐based clinical prediction using machine learning tools,” International Journal of Imaging Systems and Technology, c. 27, s. 1, ss. 89-97, 2017.

9. [9] N.A. Mayr, W.T. Yuh, J. Zheng, J.C. Ehrhardt, V.A. Magnotta, J.I. Sorosky, R.E. Pelsang, L.W. Oberley, D.H. Hussey, “Prediction of tumor control in patients with cervical cancer: analysis of combined volume and dynamic enhancement pattern by MR imaging,” American Journal of Roentgenology, c. 170, s. 1, ss. 177-182, 1998.
10. [10] M. Galar, A. Fernandez, E. Barrenechea, H. Bustince, F. Herrera, “A review on ensembles for the class imbalance problem: bagging-, boosting-, and hybrid-based approaches,” IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), c. 42, s. 4, ss. 463-484, 2012.
11. [11] P. Bühlmann, “Bagging, Boosting and Ensemble Methods,” Handbook of Computational Statistics, 2. baskı. Berlin: Springer-Verlag, 2012, ss. 985-1022.
12. [12] L. Breiman, “Random Forests,” Machine Learning, c. 45, s. 1, ss. 5–32, 2001.
13. [13] P. Geurts, D. Ernst, L. Wehenkel, “Extremely randomized trees,” Machine Learning, c. 63, s. 1, ss. 3-42, 2006.
14. [14] Y. Freund, R.E. Schapire, “A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting,” Journal of Computer and System Sciences, c. 55, s. 1, ss. 119-139, 1997.
15. [15] T. Hastie, S. Rosset, J. Zhu, H. Zou, “Multi-Class Adaboost,” Statistics and its Interface, c. 2, s. 3, ss. 349-360, 2009.
16. [16] T. Hastie, R. Tibshirani, J. Friedman, The Elements of Statistical Learning, Second Edition, New York: Springer series in statistics. 2009, böl. 10, ss. 337-384.
17. [17] F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, É. Duchesnay, “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research, c. 12, s. Oct, ss. 2825-2830, 2011.
18. [18] W. Wu, H. Zhou, “Data-Driven Diagnosis of Cervical Cancer with SVM-Based Approaches,” IEEE Access, c. 5, ss. 25189-25195, 2017.