From the ROC plot, It compares the train model vs. the valid model performance and it clearly states that train model is more efficient than valid model. Train model has more efficiency to predict that which customers would purchase the affinity card.
Scorecard Points Overall N Overall % AFFINITY_CARD has an affinity card no affinity card N % N %
BULK_PACK_DISKETTES NO PURCHASE 0.00 276.00 36.85 73.00 38.42 203.00 36.31 PURCHASE 24.00 473.00 63.15 117.00 61.58 356.00 63.69
FLAT_PANEL_MONITOR NO PURCHASE 23.00 313.00 41.79 85.00 44.74 228.00 40.79 PURCHASE 0.00 436.00 58.21 105.00 55.26 331.00 59.21 age 1: LOW - 30.5 46.00 227.00 30.31 14.00 7.37 213.00 38.10 2: 30.5 - HIGH 0.00 522.00 69.69 176.00 92.63 346.00 61.90 education 10TH 336.00 27.00 3.60 2.00 1.05 25.00 4.47 …show more content…
217.00 26.00 3.47 4.00 2.11 22.00 3.94 WIDOWED 163.00 17.00 2.27 1.00 0.53 16.00 2.86 occupation ? 109.00 40.00 5.34 6.00 3.16 34.00 6.08 ARMED-F 142.00 2.00 0.27 1.00 0.53 1.00 0.18 CLERIC. 0.00 80.00 10.68 17.00 8.95 63.00 11.27 CRAFTS 108.00 94.00 12.55 22.00 11.58 72.00 12.88 EXEC. 37.00 97.00 12.95 47.00 24.74 50.00 8.94 FARMING 13.00 18.00 2.40 4.00 2.11 14.00 2.50 HANDLER 107.00 24.00 3.20 2.00 1.05 22.00 3.94 HOUSE-S 73.00 3.00 0.40 0.00 0.00 3.00 0.54 MACHINE 108.00 57.00 7.61 8.00 4.21 49.00 8.77 OTHER 127.00 84.00 11.21 1.00 0.53 83.00 14.85 PROF. 51.00 93.00 12.42 40.00 21.05 53.00 9.48 PROTEC. 95.00 10.00 1.34 4.00 2.11 6.00 1.07 SALES 85.00 95.00 12.68 24.00 12.63 71.00 12.70 TECHSUP 53.00 19.00 2.54 4.00 2.11 15.00 2.68 TRANSP. 109.00 33.00 4.41 10.00 5.26 23.00 4.11
From the selected variable importance curve we learned that marital status was the most important variable in our model. The above cross tabulation scorecard displays the high scorecard points for the factor which influence customer to purchase an affinity card. Customer who are separated has high chance of purchasing an affinity card are married customer followed by Never married customer, divorce customer, Mabsent customer, widowed customer and the least is mar-af customer. Married customer has high chance of buying an affinity card
Scorecard Points Overall N Overall % AFFINITY_CARD has an affinity card no affinity card N % N %
BULK_PACK_DISKETTES NO PURCHASE 0.00 276.00 36.85 73.00 38.42 203.00 36.31 PURCHASE 24.00 473.00 63.15 117.00 61.58 356.00 63.69
FLAT_PANEL_MONITOR NO PURCHASE 23.00 313.00 41.79 85.00 44.74 228.00 40.79 PURCHASE 0.00 436.00 58.21 105.00 55.26 331.00 59.21 age 1: LOW - 30.5 46.00 227.00 30.31 14.00 7.37 213.00 38.10 2: 30.5 - HIGH 0.00 522.00 69.69 176.00 92.63 346.00 61.90 education 10TH 336.00 27.00 3.60 2.00 1.05 25.00 4.47 …show more content…
217.00 26.00 3.47 4.00 2.11 22.00 3.94 WIDOWED 163.00 17.00 2.27 1.00 0.53 16.00 2.86 occupation ? 109.00 40.00 5.34 6.00 3.16 34.00 6.08 ARMED-F 142.00 2.00 0.27 1.00 0.53 1.00 0.18 CLERIC. 0.00 80.00 10.68 17.00 8.95 63.00 11.27 CRAFTS 108.00 94.00 12.55 22.00 11.58 72.00 12.88 EXEC. 37.00 97.00 12.95 47.00 24.74 50.00 8.94 FARMING 13.00 18.00 2.40 4.00 2.11 14.00 2.50 HANDLER 107.00 24.00 3.20 2.00 1.05 22.00 3.94 HOUSE-S 73.00 3.00 0.40 0.00 0.00 3.00 0.54 MACHINE 108.00 57.00 7.61 8.00 4.21 49.00 8.77 OTHER 127.00 84.00 11.21 1.00 0.53 83.00 14.85 PROF. 51.00 93.00 12.42 40.00 21.05 53.00 9.48 PROTEC. 95.00 10.00 1.34 4.00 2.11 6.00 1.07 SALES 85.00 95.00 12.68 24.00 12.63 71.00 12.70 TECHSUP 53.00 19.00 2.54 4.00 2.11 15.00 2.68 TRANSP. 109.00 33.00 4.41 10.00 5.26 23.00 4.11
From the selected variable importance curve we learned that marital status was the most important variable in our model. The above cross tabulation scorecard displays the high scorecard points for the factor which influence customer to purchase an affinity card. Customer who are separated has high chance of purchasing an affinity card are married customer followed by Never married customer, divorce customer, Mabsent customer, widowed customer and the least is mar-af customer. Married customer has high chance of buying an affinity card