Artificial Intelligence Based Customer Churn Prediction Model for Business Markets

Table 4

Performance Evaluation of Distinct Runs on Proposed AICCP-TBM Method

Dataset-1
Training Size (%)	Sensitivity	Specificity	Accuracy	F-Score
K = 40	95.62	97.02	97.20	97.60
K = 50	97.00	96.44	97.87	98.40
K = 60	96.81	98.00	97.78	98.06
K = 70	96.90	97.22	97.21	96.76
K = 80	97.04	97.99	96.19	97.25
Average	96.67	97.33	97.25	97.61
Dataset-2
Number of Runs	Sensitivity	Specificity	Accuracy	F-Score
K = 40	96.61	98.09	97.81	97.88
K = 50	96.71	97.96	98.05	97.73
K = 60	97.18	96.72	97.54	97.19
K = 70	97.12	98.20	97.45	98.63
K = 80	96.97	97.53	97.63	97.07
Average	96.92	97.70	97.70	97.70
Dataset-3
Number of Runs	Sensitivity	Specificity	Accuracy	F-Score
K = 40	96.28	95.31	94.47	94.42
K = 50	95.81	94.40	96.57	93.21
K = 60	97.01	93.95	96.36	92.40
K = 70	94.94	95.31	92.34	93.21
K = 80	95.38	94.86	91.92	93.20
Average	95.88	94.77	94.33	93.29