Table 3 Feature importance analysis
From: Machine learning algorithm for early detection of end-stage renal disease
Feature | Feature importance |
|---|---|
Age | 0.030 |
CKD stage | 0.018 |
Hypertensive crisis events per year | 0.016 |
Recently diagnosed hypertension | 0.013 |
Total drug prescriptions per year | 0.010 |
Total cost of outpatient specialist visits per year | 0.007 |
Annual medication costs | 0.006 |
Hypertensive nephropathy | 0.006 |
Recently diagnosed hyperlipidemia | 0.006 |
Time gap between last CKD stage diagnosis to most recent | 0.004 |
Number of urinalysis tests per year | 0.004 |
Ever diagnosis of hypertension | 0.004 |
Total cost of ER and inpatient visits per year | 0.003 |
Total annual claims costs | 0.003 |
Acute kidney injury events per year | 0.002 |
Anemia of CKD | 0.002 |
Recently diagnosed diabetes | 0.002 |