Makine Öğrenimi Modelleri ile Kırılganlık Seviyesi Sınıflandırma ve Veri Tipi Etkisi

Yıl 2025, Cilt: 25 Sayı: 5, 1106 - 1113, 01.10.2025

Mehmet Halit Özalp , Didem Ölçer

https://doi.org/10.35414/akufemubid.1595455

Öz

Günümüzde kırılganlık, yaşlanan nüfus ile önemli bir sorun haline gelmektedir. Kırılganlık erken teşhis edilmesi ve uygun müdahaleler ile azaltılabilen veya tersine çevrilebilen geriatrik bir sendromdur. Bu nedenle bu çalışmada, KNN (K-Nearest Neighbors), Naïve Bayes, XGBoost (Extreme Gradient Boosting), RF (Random Forests), MLP (Multi-Layer Perceptron), 1D-CNN (1D Convolutional Neural Network) modelleri ile, STAIR testi sırasında toplanan EKG, üç eksenli ivme ölçer ve bu iki sensor verilerinin birleşiminden elde edilen üç ayrı veri seti üzerinde zaman alanı öznitelikleri kullanılarak kırılganlık sınıflandırma problemine çözüm aranmaktadır. Çalışmanın sonucunda RF ve XGBoost modellerinin en başarılı sonuçları verdiği, ayrıca EKG verilerinin kırılganlık tahmininde kullanımının akselerometre verilerinin kullanımından daha yüksek doğrulukta tahminleme yapılabildiği görülmüştür. Ayrıca Kırılganlık tahmininde daha doğru sonuçlar veren verilerin takibi ile birlikte, daha kesin sonuçlarla kişiler, kırılganlık durumlarını izleyebilecek ve sadece gerekli olduğunda bir uzman görüşüne danışacaktır.

Anahtar Kelimeler

EKG , İvmeölçer , Kırılganlık Sınıflandırma , Makine Öğrenmesi , Öznitelik Çıkarımı , STAIR testi

Frailty Level Classification with Machine Learning Models and Data Type Effect

Öz

Today, frailty is becoming an important problem with the ageing population. Frailty is a geriatric syndrome that can be reduced or reversed by early detection and appropriate interventions. Therefore, in this study, KNN (K-Nearest Neighbors), Naïve Bayes, XGBoost (Extreme Gradient Boosting), RF (Random Forests), MLP (Multi-Layer Perceptron), 1D-CNN (1D Convolutional Neural Network) models are used to solve the frailty classification problem by using time domain features on three separate data sets obtained from the combination of ECG, triaxial accelerometer and these two sensor data collected during the STAIR test. As a result of the study, it is seen that RF and XGBoost models give the most successful results, and the use of ECG data in fragility prediction can be predicted with higher accuracy than the use of accelerometer data. In addition, with the follow-up of data that give more accurate results in frailty prediction, people will be able to monitor their frailty status with more precise results and consult an expert opinion only when necessary.

Anahtar Kelimeler

Accelerometer , ECG , Feature Extraction , Frailty Classification , Machine Learning , STAIR test

Kaynakça

• Alan, A., ve Karabatak, M., 2020. Veri Seti - Sınıflandırma İlişkisinde Performansa Etki Eden Faktörlerin Değerlendirilmesi. Fırat Üniversitesi Mühendislik Bilimleri Dergisi, 32(2), 531-540. https://doi.org/10.35234/fumbd.738007
• Aydemir, Ö., 2022. Dengesiz veri kümelerinin sınıflandırılmasında poligon alan metriğinin sınıflandırıcı performans değerlendirilmesi için kullanılması. Yüzüncü Yıl Üniversitesi Fen Bilimleri Enstitüsü Dergisi, 27(2), 194–205. https://doi.org/10.53433/yyufbed.1066340
• Aygör, H., Fadıloğlu, Ç., Şahin, S., Aykar, F., and Akçiçek, F., 2018. Validation of Edmonton Frail Scale into elderly Turkish population. Archives of Gerontology and Geriatrics, 76, 133-137. https://doi.org/10.1016/j.archger.2018.02.003
• Barrash, S., Shen, Y., and Giannakis, G.B., 2019. Scalable and Adaptive kNN for Regression Over Graphs. IEEE CAMSAP 2019, 241-245. https://doi.org/10.1109/CAMSAP45676.2019.9022509
• Bozkurt Keser, S., ve Keskin, K., 2023. Kalp Yetmezliği Hastalarının Sağ Kalım Tahmini: Sınıflandırmaya Dayalı Makine Öğrenmesi Algoritmalarının Bir Uygulaması. Afyon Kocatepe Üniversitesi Fen ve Mühendislik Bilimleri Dergisi, 23, 362-369. https://doi.org/10.35414/akufemubid.1033377
• Breiman, L., 2001. Random Forests. Machine Learning, 45, 5-32. https://doi.org/10.1023/A:1010933404324
• Chen, T., and Guestrin, C., 2016. XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785-784. https://doi.org/10.1145/2939672.2939785
• Chen, Y., Yu, Y., Yang, D., Zhang, W., Kouritas, V., and Chen, X., 2024. Developing and validating machine learning-based prediction models for frailty occurrence in those with chronic obstructive pulmonary disease. Journal of Thoracic Disease, 16(4), 2482-2498. https://doi.org/10.21037/jtd-24-416
• Erdaş, Ç.B., Atasoy, I., Açıcı, K., and Oğul, H., 2016. Integrating Features for Accelerometer-based Activity Recognition. Procedia Computer Science, 98, 522-527. https://doi.org/10.1016/j.procs.2016.09.070
• Erdaş, Ç.B., and Ölçer, D., 2022. Prediction of Frailty Grade Using Machine Learning Models. In 2022 Medical Technologies Congress (TIPTEKNO), Antalya, Türkiye, 1-4. https://doi.org/10.1109/TIPTEKNO56568.2022.9960172
• Erdaş, Ç.B., Sümer, E., and Kibaroğlu, S., 2021. Neurodegenerative disease detection and severity prediction using deep learning approaches. Biomedical Signal Processing and Control, 70, 103069. https://doi.org/10.1016/j.bspc.2021.103069
• Hoogendijk, E., Afilalo, J., Ensrud, K., Kowal, P., Onder, G., and Fried, L., 2019. Frailty: implications for clinical practice and public health. The Lancet, 394(10206), 1365-1375. https://doi.org/10.1016/s0140-6736(19)31786-6
• Kikuchi, H., Inoue, S., Amagasa, S., Fukushima, N., Machida, M., Murayama, H., Fujiwara, T., Chastin, S., Owen, N., & Shobugawa, Y. (2020). Associations of older adults' physical activity and bout-specific sedentary time with frailty status: Compositional analyses from the NEIGE study. Experimental Gerontology. https://doi.org/10.1016/j.exger.2020.111149
• Kiranyaz, S., Ince, T., and Gabbouj, M., 2015. Real-Time Patient-Specific ECG Classification by 1D Convolutional Neural Networks. IEEE Transactions on Biomedical Engineering, 63(3), 664-675. https://doi.org/10.1109/TBME.2015.2468589
• Leme, D.E.C., and Oliveira, C., 2023. Machine Learning Models to Predict Future Frailty in Community-Dwelling Middle-Aged and Older Adults: The ELSA Cohort Study. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 78(11), 2176–2184. https://doi.org/10.1093/gerona/glad127
• Mizuguchi, Y., Nakao, M., Nagai, T. (2024). Machine learning–based gait analysis to predict clinical frailty scale in elderly patients with heart failure. European Heart Journal - Digital Health, 5(2), 152–162. https://doi.org/10.1093/ehjdh/ztad082
• Panhwar, Y., Naghdy, F., Naghdy, G., Stirling, D., and Potter, J., 2019. Assessment of frailty: a survey of quantitative and clinical methods. BMC Biomedical Engineering, 1-20. https://doi.org/10.1186/s42490-019-0007-y
• Razjouyan, J., Naik, A. D., Horstman, M. J., Kunik, M. E., Amirmazaheri, M., Zhou, H., Sharafkhaneh, A., & Najafi, B. (2018). Wearable sensors and the assessment of frailty among vulnerable older adults: An observational cohort study. Sensors, 18(5), 1336. https://doi.org/10.3390/s18051336
• Rolfson, D., Majumdar, S., Tsuyuki, R., Tahir, A., and Rockwood, K., 2006. Validity and reliability of the Edmonton Frail Scale. Age and Ageing, 35(5), 526-529. https://doi.org/10.1093/ageing/afl041
• Sano, A.V.D., Stefanus, A.A., Madyatmadja, E.D., Nindito, H., Purnomo, A., and Sianipar, C.P.M., 2023. Proposing a visualized comparative review analysis model on tourism domain using Naïve Bayes classifier. Procedia Computer Science, 227, 482-489. https://doi.org/10.1016/j.procs.2023.10.549
• Sokas, D., Butkuvienė, M., Tamulevičiūtė-Prascienė, E., Beigienė, A., Kubilius, R., Petrėnas, A., and Paliakaitė, B., 2022. Wearable-based signals during physical exercises from patients with frailty after open-heart surgery (version 1.0.0). PhysioNet. https://doi.org/10.13026/mp8k-7p27
• Szczepanowski, R., Uchmanowicz, I., Pasieczna, A.H., Sobecki, J., Katarzyniak, R., Kołaczek, G., Lorkiewicz, W., Kędras, M., Dixit, A., Biegus, J., Wleklik, M., Gobbens, R.J.J., Hill, L., Jaarsma, T., Hussain, A., Barbagallo, M., and Morabito, F.C., 2024. Application of machine learning in predicting frailty syndrome in patients with heart failure. Advances in Clinical and Experimental Medicine, 33(3), 309-315. https://doi.org/10.17219/acem/184040
• Wang, Y., Wei, S., Zuo, R., et al., 2024. Automatic and real-time tissue sensing for autonomous intestinal anastomosis using hybrid MLP-DC-CNN classifier-based optical coherence tomography. Biomedical Optics Express, 15(4), 2543-2557. https://doi.org/10.1364/BOE.521652