Ferritin Level Prediction in Patients with Chronic Kidney Disease Using Cluster Centers on Fuzzy Subtractive Clustering

Abstract

It is important to know about iron reserves in patients on hemodialysis who have chronic kidney disease (CKD). Early detection of iron insufficiency or raised serum iron levels is crucial. Ferritin levels are one thing that can be used to cool this. Regretfully, ferritin testing is still seen as a highly costly procedure. Therefore, utilizing straightforward and affordable variables including height, weight, blood pressure, the duration of hemodialysis, history of comorbidities, and Hb levels before and after hemodialysis, this study predicts ferritin levels. The cluster center is used to help in ferritin level prediction. Due to the wide diversity of the sample data, the clustering technique is applied for clustering. Fuzzy subtractive clustering (FSC) was used to adaptively categorize 50 patient states using the dense concept. After clustering, we wound up with eight final clusters that had an accept ratio of 0.75, a reject ratio of 0.25, and an influence range of 0.5. The blood pressure variable has the strongest link with ferritin levels, according to the correlation coefficient. The mean degree of agreement between ferritin levels in the real and predicted samples was 62.53%. After evaluating nine sets of test data, the average similarity value was 83.74%. When data is clustered using the K-Means approach, the application of cluster centers yields a result of 50.91%; this result is significantly higher. This study’s limitation is that it is unable to identify the ideal cluster in the presence of numerous outliers. Consequently, it is necessary to conduct additional study while taking the ideal number of clusters into consideration.