| Autor: |
Klinsega Jeberson, Raghav Yadav, Lordwin Jeyakumar, Manish Kumar |
| Rok vydání: |
2020 |
| Předmět: |
|
| Zdroj: |
Machine Intelligence and Signal Processing ISBN: 9789811513657 |
| DOI: |
10.1007/978-981-15-1366-4_22 |
| Popis: |
Chronic kidney disease (CKD) largely affects people worldwide and is now seen as a common health threat attributable to its escalating prevalence and huge costs associated with dialysis and transplantation, which are therapies for advanced CKD. Due to various factors, the detection of the condition in patients is being delayed. Also, the markers like serum creatinine and albumin which are commonly used for CKD screening in clinical practice seem to be inadequate. In this study, an approach combining J48 and LADTree has been proposed. The outcomes indicate that the proposed approach performed exceptionally well with perfect accuracy in cross validation. The proposed model exhibited better performance than doctor 1 and doctor 2 on a test set containing 100 records in an additional validation process with a performance difference of 10% and 20%, respectively. The model may be adopted to construct decision support system that enables inexperienced physicians to accurately detect CKD using few laboratory values. Low-cost gadgets based on the proposed model may be developed which may extend the reach of health-care professionals outside the clinic and therefore aid timely diagnosis of the condition in large proportion of population at risk. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
