Research Article
Prediction of Repeat Customers on E-Commerce Platform Based on Blockchain
Table 3
The accuracy and AUC values of five types of ensemble prediction models.
| Model | Training set | Test set | Accuracy | AUC | Accuracy | AUC | Mean | Variance | Mean | Variance | Mean | Variance | Mean | Variance |
| XGBoost | 0.7762 | 0.0412 | 0.6752 | 0.0408 | 0.7751 | 0.0396 | 0.6744 | 0.0419 | LightGBM | 0.7930 | 0.0398 | 0.6878 | 0.0406 | 0.7908 | 0.0390 | 0.6875 | 0.0404 | DeepGBM | 0.8664 | 0.0361 | 0.7022 | 0.0356 | 0.8576 | 0.0371 | 0.6991 | 0.0366 | Stacking | 0.9080 | 0.0336 | 0.7027 | 0.0319 | 0.8943 | 0.0338 | 0.7023 | 0.0339 | Vote-stacking | 0.9128 | 0.0291 | 0.7053 | 0.0302 | 0.9014 | 0.0334 | 0.7050 | 0.0302 |
|
|