Deep Neural Networks and an Application in Health Sciences
Yıl: 2021 Cilt: 28 Sayı: 4 Sayfa Aralığı: 609 - 614 Metin Dili: İngilizce DOI: 10.5505/vtd.2021.83707 İndeks Tarihi: 22-09-2022
Deep Neural Networks and an Application in Health Sciences
Öz: Objective: Because there is more than one hidden layer between the input and output layers in the neural network algorithm, it is called "Deep Neural Networks". In the study, the Deep Neural Networks algorithm; different input (number of layers, epoch, error rate) and evaluation of the performance of the model being practiced an application is intended. Materials and Methods: The most important feature that distinguishes the deep neural network method from the classical neural networks is the number of layers that provide good results in complex problems. Wart treatment results of patients who used immunotherapy were used as data set in the study. Results: According to this; the simple-layer (1 hidden layer) artificial neural network model was classified with 87.5% overall accuracy and 29.74% MAPE, whereas the deep neural network model was classified with 99.8% overall accuracy and 25.19% MAPE ratio. The study showed that the model of deep neural networks had a higher accuracy rate. Conclusion: As a result of the application performed in this study; it is seen that the multi-layered (deep) neural network model provides classification with higher Accuracy and lower error rate than the single layer (classical) neural network model. In other words, according to the results of this study; it has been found that the deep neural network model has a better (optimum) classification rate than the classical neural network model.
Anahtar Kelime: Derin Sinir Ağlarıve Sağlık Bilimlerinde Bir Uygulama
Öz: Amaç: Yapay sinir ağları algoritmasında giriş ve çıkış katmanları arasında birden fazla gizli katman bulunduğundan, “Derin Sinir Ağları” (Deep Neural Network) olarak adlandırılır. Klasik bir yapay sinir ağında bulunan nöronların birbirleriyle ilişkileri yoktur ve bilgi ancak giriş katmanından çıkış katmanına doğru aktarılır. Derin sinir ağlarında ise artarda gelen iki katmanda nöronlar birbirlerini çeşitli aktivasyon değerleriyle etkilemektedir. Her katmanın modele etkisi ve dolayısıyla her katmandaki nöronun modele etkisi bulunmaktadır. Çalışmada, Derin Sinir Ağları algoritması; farklı girdi (katman sayısı, döngü, hata oranı) ve uygulamalı modelin performansının değerlendirilmesi amaçlanmaktadır. Gereç ve Yöntem: Derin sinir ağı yöntemini klasik sinir ağlarından ayıran en önemli özellik, karmaşık problemlerde iyi sonuçlar veren katman sayısıdır. Çalışmada veri seti olarak immünoterapi kullanan hastaların siğil tedavisi sonuçları verisi kullanıldı. Bulgular ve Sonuç: Basit katmanlı (bir gizli katman) yapay sinir ağı modelinde %87.5 genel doğruluk ve %29.74 MAPE ile sınıflandırılırken, derin sinir ağı modeli %99.8 genel doğruluk ve %25.19 MAPE oranı ile sınıflandırılmıştır. Çalışma, derin sinir ağları modelinin daha yüksek doğruluk oranına sahip olduğunu göstermiştir.
Anahtar Kelime: Belge Türü: Makale Makale Türü: Araştırma Makalesi Erişim Türü: Erişime Açık
- 1. Naylor MF, Neldner KH, Yarbrough GK, Rosio TJ, Iriondo M, Yeary J. Contact immunotherapy of resistant warts. J Am Acad Dermatol 1988; 19(4):679-683.
- 2. Zur Hausen H,de Villiers EM. Human papillomaviruses. Annu Rev Microbiol 1994; 48(1):427-447.
- 3. Chandrashekar L. Intralesional immunotherapy for the management of warts. Indian J Dermatol Venereol Leprol 2011; 77(3):261-262.
- 4. Khozeimeh F, Alizadehsani R, Roshanzamir M, Khosravi A, Layegh P, Nahavandi S. An expert system for selecting wart treatment method. Comput Biol Med 2017; 81;2(1):167-175.
- 5. UCI. 2019. University of California, Irvine. https://archive.ics.uci.edu/ml/datasets/I mmunotherapy+Dataset Accessed: 15.06.2019.
- 6. Khozeimeh F, Jabbari Azad F, Mahboubi Oskouei Y, Jafari M, Tehranian S, Alizadehsani R, et al. Intralesional immunotherapy compared to cryotherapy in the treatment of warts. Int J Dermatol 2017; 56(4):474-478.
- 7. Rende FŞ, Bütün G, Karahan Ş. Derin Öğrenme Algoritmalarında Model Testleri: Derin Testler. 10. Ulusal Yazıl. Müh. Semp 24-26 Ekim 2016; 54-59, Çanakkale.
- 8. Wang Y, Mao H, Yi Z. Protein secondary structure prediction by using deep learning method. Knowledge-Based Systems 2017; 118:115-123.
- 9. LeCun Y, Bengio Y, Hinton G. Deep learning, Nature 2015; 521:436-444.
- 10. Brownlee J. 2016. How to Grid Search Hyperparameters for Deep Learning Models in Python with Keras, https://machinelearningmastery.co m/grid-search-hyperparameters-deeplearning- models-python-keras/, Accessed: 21.04.2019.
- 11. Çarkacı N. 2018. Derin Öğrenme Uygulamalarında En Sık kullanılan Hiper parametreler. https://medium.com/deeplearning- turkiye/derin-ogrenmeuygulamalarinda- en-sik-kullanilan-hiperparametreler- ece8e9125c4 Accessed: 19.04.2019.
- 12. Schmidhuber J. Deep learning in neural networks: An overview. Neur Netw 2015; 61: 85-117.
- 13. Güzel Y. 2018. Yapay Zeka Ders Notları 05. Çok Katmanlı Yapay Sinir Ağı. https://medium.com/@yasinguzel/yapayzeka- ders-notlar%C4%B1-5-%C3%A7okkatmanl% C4%B1-yapay-sinira% C4%9F%C4%B1-4c6af075e1fe Accessed: 02.05.2019.
- 14. Güzel K. 2018. Geri Yayılımlı Çok Katmanlı Yapay Sinir Ağları-1. https://kadirguzel.medium.com/geriyay% C4%B1l%C4%B1ml%C4%B1- %C3%A7ok-katmanl%C4%B1-yapaysinir- a%C4%9Flar%C4%B1-1- 47daa3856247 Accessed: 15.04.2019.
| APA | ELASAN S (2021). Deep Neural Networks and an Application in Health Sciences. , 609 - 614. 10.5505/vtd.2021.83707 |
| Chicago | ELASAN Sadi Deep Neural Networks and an Application in Health Sciences. (2021): 609 - 614. 10.5505/vtd.2021.83707 |
| MLA | ELASAN Sadi Deep Neural Networks and an Application in Health Sciences. , 2021, ss.609 - 614. 10.5505/vtd.2021.83707 |
| AMA | ELASAN S Deep Neural Networks and an Application in Health Sciences. . 2021; 609 - 614. 10.5505/vtd.2021.83707 |
| Vancouver | ELASAN S Deep Neural Networks and an Application in Health Sciences. . 2021; 609 - 614. 10.5505/vtd.2021.83707 |
| IEEE | ELASAN S "Deep Neural Networks and an Application in Health Sciences." , ss.609 - 614, 2021. 10.5505/vtd.2021.83707 |
| ISNAD | ELASAN, Sadi. "Deep Neural Networks and an Application in Health Sciences". (2021), 609-614. https://doi.org/10.5505/vtd.2021.83707 |
| APA | ELASAN S (2021). Deep Neural Networks and an Application in Health Sciences. Van Tıp Dergisi, 28(4), 609 - 614. 10.5505/vtd.2021.83707 |
| Chicago | ELASAN Sadi Deep Neural Networks and an Application in Health Sciences. Van Tıp Dergisi 28, no.4 (2021): 609 - 614. 10.5505/vtd.2021.83707 |
| MLA | ELASAN Sadi Deep Neural Networks and an Application in Health Sciences. Van Tıp Dergisi, vol.28, no.4, 2021, ss.609 - 614. 10.5505/vtd.2021.83707 |
| AMA | ELASAN S Deep Neural Networks and an Application in Health Sciences. Van Tıp Dergisi. 2021; 28(4): 609 - 614. 10.5505/vtd.2021.83707 |
| Vancouver | ELASAN S Deep Neural Networks and an Application in Health Sciences. Van Tıp Dergisi. 2021; 28(4): 609 - 614. 10.5505/vtd.2021.83707 |
| IEEE | ELASAN S "Deep Neural Networks and an Application in Health Sciences." Van Tıp Dergisi, 28, ss.609 - 614, 2021. 10.5505/vtd.2021.83707 |
| ISNAD | ELASAN, Sadi. "Deep Neural Networks and an Application in Health Sciences". Van Tıp Dergisi 28/4 (2021), 609-614. https://doi.org/10.5505/vtd.2021.83707 |
