Research Article

Implementation of Predictive Data Mining Techniques for Identifying Risk Factors of Early AVF Failure in Hemodialysis Patients

Table 3

Extracted rules, analysis with side of AVF.

Rule model

If location = brachial and DM = no, then yes (30/2)
If location = radial and DM = yes, then no (24/45)
If dm = yes and Hgb = range2 [8.450–9.950], then yes (5/0)
If Hgb = range3 [9.950– ] and htn = yes, then yes (11/5)
If age = range1 [ –49.500] and Hgb = range3 [9.950– ], then yes (7/1)
If age = range2 [49.50–64.50] and sex = female, then no (0/3)
If Hgb = range3 [9.950– ], then no (0/4)
If htn = no and Hgb = range1 [ –8.450], then yes (9/4)
If age = range2 [49.500–64.500] and DM = no, then no (1/5)
If htn = yes and sex = female, then yes (4/2)
If sex = female and htn = no, then no (1/2)
If age = range3 [64.500– ] and Hgb = range1 [ –8.450], then no (2/5)
If htn = yes and age = range3 [64.500– ], then yes (2/0)
If Hgb = range1 [ –8.450] and DM = yes, then yes (1/0)
If Hgb = range2 [8.450–9.950] and htn = yes, then no (1/2)
If htn = yes and sex = male, then yes (3/2)
If sex = male, then no (5/5)
Correct: 143 out of 193 training examples