BioMed Research International

Research Article

Effect of Missing Data Imputation on Deep Learning Prediction Performance for Vesicoureteral Reflux and Recurrent Urinary Tract Infection Clinical Study

Table 1

The variables used in deep learning and multiple imputation techniques.


Clinical variables	Laboratory variables	USG variables

Diagnosis_(VUR/rUTI)	ud-density_(c)	USG-R-grade _{(ordinal:0,1,2)}
Sex_{(cat: male/female)}	b-leukocyte_(c)	USG-L-grade_{(ordinal: 0,1,2)}
Age_(c)	ud-nitrite_(cat:Y/N)	USG-R/L hydronephrosis _{(cat: Y/N)}
Fever_{(cat: Y/N)}	ud-l.esterase_{(cat: Y/N)}	USG-bladder wall thickening_{(cat: Y/N)}
Emesis_(catty/N)	ud-protein_{(cat: Y/N)}	USG-bladder diverticulum_{(cat: Y/N)}
Incontinence_{(cat: Y/N)}	us-erythrocyte_{(cat: Y/N)}	USG-ureter dilatation R/L _{(cat: Y/N)}
Stomachache_{(cat: Y/N)}	us-leukocyte_{(cat: Y/N)}
Urgency_{(cat: Y/N)}	ud-leukocyte_{(cat: Y/N)}
Frequent urination_{(cat: Y/N)}	us-bacteria _{(cat: Y/N)}
Dysuria_{(cat: Y/N)}	ud-erythrocyte_(cat)
Restlessness_{(cat: Y/N)}	b-thrombocyte_(c)
Anorexia_{(cat: Y/N)}	b-urea_(c)
UTI in history_{(cat: Y/N)}	b-creatinine_(c)
Prolonged neonatal jaundice_{(cat: Y/N)}
Scar_{(cat: Y/N)}

All categorical variables are defined as binary (cat: Y/N, yes/no, and sex, cat: male/female). c: continuous variable; cat: categorical variable; rUTI: recurrent urinary tract infection; ud: urine dipstick; us: urine sediment; USG: ultrasonography; b: blood; R: right; L: left; u-le: urine-leukocyte esterase.