Penerapan data mining pada prilaku konsumen pada produk x

Data

Algoritma

Pohon Keputusan ( Decissio Tree )

Hasil

Age > 31.500
| Gender = female: no {yes=0, no=324}
| Gender = male
| | Payment Method = cash
| | | Age > 35.500: no {yes=0, no=109}
| | | Age ≤ 35.500: yes {yes=8, no=0}
| | Payment Method = cheque
| | | Age > 35.500: no {yes=0, no=23}
| | | Age ≤ 35.500: yes {yes=5, no=0}
| | Payment Method = credit card
| | | Age > 72.500: no {yes=0, no=40}
| | | Age ≤ 72.500: yes {yes=221, no=0}
Age ≤ 31.500
| Payment Method = cash
| | Gender = female: no {yes=0, no=20}
| | Gender = male: yes {yes=44, no=0}
| Payment Method = cheque
| | Gender = female: no {yes=0, no=10}
| | Gender = male: yes {yes=6, no=0}
| Payment Method = credit card: yes {yes=189, no=1}

Kinerja Algoritma

PerformanceVector:
accuracy: 99.60% +/- 0.66% (mikro: 99.60%)
ConfusionMatrix:
True: yes no
yes: 473 4
no: 0 523
precision: 100.00% +/- 0.00% (mikro: 100.00%) (positive class: no)
ConfusionMatrix:
True: yes no
yes: 473 4
no: 0 523
recall: 99.23% +/- 1.27% (mikro: 99.24%) (positive class: no)
ConfusionMatrix:
True: yes no
yes: 473 4
no: 0 523
AUC (optimistic): 0.998 +/- 0.003 (mikro: 0.998) (positive class: no)
AUC: 0.897 +/- 0.199 (mikro: 0.897) (positive class: no)
AUC (pessimistic): 0.992 +/- 0.013 (mikro: 0.992) (positive class: no)

Dengan kinerja algortima 99,6 % maka algortima bisa di terapkan dengan sangat baik