Publications by authors named "Seoksu Hong"
Diagnostics (Basel)
June 2020
Article Synopsis
- The study compares the classification effectiveness of various statistical models on a dataset related to chronic kidney disease (CKD) using the National Health Insurance Service database in Korea.
- It evaluates different machine learning methods, including multinomial logistic regression, ordinal logistic regression, random forest, and autoencoder, focusing on the accurate classification of CKD stages based on glomerular filtration rate (GFR).
- Results reveal that the autoencoder model outperforms others in correctly classifying CKD stages, particularly when considering multiple performance metrics like accuracy, sensitivity, and precision, especially in situations with imbalanced data.
