br Conclusion br Cervical cancer can hide itself
Cervical cancer can hide itself for a long time, but shortly after the treatment, it may be helpful if the doctor can predict the re-currence of disease based on several variables. Doctors are always expected to make the right decision. Many years of clinical work and experiments have to be done, to be able to make the right decision, but still faults are always possible. To better predict the disease, many investigators have tried to identify risk factors such as tumor size, depth, and the likelihood of regrowth affecting the
disease. In this study, It was determined the disease to the data set by applied dimension reduction.
In this study, in order to classify cervical cancer differently from the studies in the literature, the other machine learning methods as well as the stacked autoencoder, which are deep learning so-lutions, were used for the first time. One of the major disadvan-tages of many machine learning methods is the di culty in dimen-sion reduction; however, this problem is easily eliminated with the deep learning. The stacked autoencoder model we used in the study works as a classifier with high accuracy by eliminating un-necessary attributes to reduce data dimension. It is seen that the proposed model in the study achieves better success rates in clas-sifying cervical cancer data compared to the other machine learn-ing methods and kPCA dimension MCC950 methods. The train-ing time performance of the proposed model in the study is worse than the other methods due to the large amount of time spent in reducing the dimension and the number of samples used in the training. This situation is ignored because a single sample will be given to the proposed model during the test. For this reason, the stacked autoencoder model can be used as an alternative method in healthcare decision support systems.
Kemal Adem – Conceptualization; Formal analysis; Methodol-ogy; Software; Validation; Writing - original draft; Writing - re-view & editing.
Serhat Kiliçarslan – Conceptualization; Investigation; Software; Writing - review & editing.
Onur Cömert – Formal analysis; Methodology; Validation; Visu-alization & Writing - original draft.
Almotiri, J., Elleithy, K., & Elleithy, A. (2017). Comparison of autoencoder and Princi-pal Component Analysis followed by neural network for e-learning using hand-written recognition. In Systems, applications and technology conference (LISAT), 2017 IEEE long island (pp. 1–5). IEEE.
Badem, H., Caliskan, A., Basturk, A., & Yuksel, M. E. (2016). Yıgınlanmıs Özdevinimli kodlayici ileinsan aktivitelerinin siniflandirilmasi classification of human activity by using a coenzymes stacked autoencoder. TIPTEKNO2016, Antalya.
Ceylan, Z., & Pekel, E. (2017). Comparison of multi-label classification methods for prediagnosis of cervical cancer. International Journal of Intelligent Systems and Applications in Engineering, 5(4), 232–236.
Chapelle, O., & Vapnik, V. (2000). Model selection for support vector machines. Ad-vances in Neural Information Processing Systems, 230–236.
Fatlawi, H. K. (2017). Enhanced classification model for cervical cancer data set based on cost sensitive classifier. International Journal of Computer Techniques, 4(4), 115–120.
(2017). Cancer statistics Turkey. Turkish Public Health Institution. http://kanser.
Kurniawati, Y. E., Permanasari, A. E., & Fauziati, S. (2016). Comparative study on data mining classification methods for cervical cancer prediction using pap smear results. In Biomedical engineering (IBIOMED), international conference (pp. 1–5). http://ieeexplore.ieee.org/document/7869827.
image analysis in multispectral system for cervical cancer diagnostic. In Open innovations association (FRUCT), 2017 20th conference of (pp. 345–351). IEEE.
cations. Singapore: Hackensack.
Sharma, S. (2016). Cervical cancer stage prediction using decision tree approach of machine learning. International Journal of Advanced Research in Computer and Communication Engineering, 5(4), 345–348. doi:10.17148/IJARCCE.2016.5488.