Comparative Analysis of Machine Learning Algorithms for Classifying Postoperative Survival Time in Lung Cancer Patients
VII. International Applied Statistics Congress (UYIK – 2026), İstanbul, Türkiye, 11 - 13 Mayıs 2026, sa.1397, ss.446-449, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.446-449
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Ondokuz Mayıs Üniversitesi Adresli: Evet
Özet
Abstract
Lung cancer is a major public health problem responsible for a significant number of deaths worldwide. According to the International Agency for Research on Cancer (IARC), in 2022 lung cancer was the most commonly diagnosed cancer globally with 2.5 million new cases, accounting for 12.4% of all new cancer cases. If incidence and mortality rates remain constant at 2022 levels, the burden of lung cancer is expected to rise to 4.62 million new cases and 3.55 million deaths by 2050. These projections highlight the increasing burden of cancer, its disproportionate impact on disadvantaged populations, and the urgent need to address global cancer inequalities. Accurate prediction of post-treatment life expectancy is critically important both for improving patients’ quality of life and for effective healthcare planning. The dataset used in this study was retrospectively collected between 2007 and 2011 by the Institute of Tuberculosis and Lung Diseases in Warsaw. Based on the assumption that traditional methods applied to limited datasets may be insufficient in terms of accuracy and consistency, this study compares the performance of two widely used machine learning algorithms for such data: Artificial Neural Networks (ANN) and K-Nearest Neighbors (KNN). To obtain reliable and valid results, particular attention was given to data preprocessing, feature selection, model training, and parameter optimization. The performance of the algorithms was evaluated comparatively to determine the most suitable model for survival time classification. The findings indicate that both ANN and KNN algorithms successfully capture the multidimensional clinical patterns affecting postoperative survival time. In conclusion, the integration of machine learning techniques into clinical decision support systems can optimize patient management by increasing prediction accuracy. However, methodological rigor and appropriate validation processes are of critical importance to prevent potential inconsistencies. This study demonstrates the necessity of comparative evaluation of algorithms in developing reliable models for predicting postoperative survival time in lung cancer patients.
Keywords: Lung Cancer, Machine Learning, Postoperative Survival Time, ANN, KNN, Clinical Decision Support Systems