Research Article
A New Approach for Construction of Geodemographic Segmentation Model and Prediction Analysis
Table 8
Confusion matrix of different classification techniques.
| | Model | Actual class | Actual prediction | | Nonchurners | Churners |
| | CART decision tree | Nonchurners | 645/740 (87.16%) | 95/740 (12.84%) | | Churners | 146/260 (56.15%) | 114/260 (43.85%) | | Random forest | Nonchurners | 698/740 (94.32%) | 42/740 (5.68%) | | Churners | 153/260 (58.8%) | 107/260 (41.15%) | | Gradient boost | Nonchurners | 716/740 (96.76%) | 24/740 (3.24%) | | Churners | 158/260 (60.77%) | 102/260 (39.23%) | | AdaBoost | Nonchurners | 704/740 (95.14%) | 36/740 (4.86%) | | Churners | 158/260 (60.77%) | 102/260 (39.23%) | | Extra trees | Nonchurners | 703/740 (95.0%) | 37/740 (5.0%) | | Churners | 162/260 (62.31%) | 98/260 (37.69%) | | SVM | Nonchurners | 721/740 (97.43%) | 19/740 (2.57%) | | Churners | 171/260 (65.77%) | 89/260 (34.23%) | | Artificial neural network | Nonchurners | 707/740 (95.54%) | 33/740 (4.46%) | | Churners | 158/260 (60.77%) | 102/260 (39.23%) | | Naïve Bayes | Nonchurners | 700/740 (94.59%) | 40/740 (5.41%) | | Churners | 189/260 (72.69%) | 71/260 (27.31%) | | kNN | Nonchurners | 722/740 (97.56%) | 18/740 (2.44%) | | Churners | 173/260 (66.54%) | 87/260 (33.46%) |
|
|
