Skip to main content

Table 3 Performance, mean and (standard deviation), of churn prediction models for the four feature groups using the 3-day dataset

From: Social networks for enhanced player churn prediction in mobile free-to-play games

Classifier

Benchmark

Benchmark + Friends

Precision

Recall

\({\mathbf{F}}_{1}\)

AUC

Precision

Recall

\(F_1\)

AUC

DecisionTree

0.688 (0.017)

0.797 (0.033)

0.738 (0.018)

0.774 (0.018)

0.692 (0.019)

0.880 (0.021)

0.774 (0.008)

0.743 (0.014)

KNN

0.538 (0.036)

0.704 (0.185)

0.599 (0.073)

0.545 (0.058)

0.537 (0.036)

0.704 (0.185)

0.599 (0.073)

0.545 (0.058)

RandomForest

0.712 (0.015)

0.840 (0.015)

0.771 (0.014)

0.812 (0.014)

0.714 (0.015)

0.834 (0.016)

0.769 (0.012)

0.812 (0.012)

XGBoost

0.705 (0.008)

0.807 (0.022)

0.752 (0.011)

0.797 (0.016)

0.702 (0.017)

0.799 (0.055)

0.747 (0.029)

0.796 (0.019)

SGD

0.622 (0.068)

0.578 (0.288)

0.546 (0.164)

0.592 (0.042)

0.614 (0.069)

0.567 (0.332)

0.516 (0.218)

0.585 (0.047)

Classifier

Benchmark + Similarity

Benchmark + Friends + Similarity

Precision

Recall

\({\mathbf{F}}_{1}\)

AUC

Precision

Recall

\({\mathbf{F}}_{1}\)

AUC

DecisionTree

0.692 (0.019)

0.880 (0.021)

0.774 (0.008)

0.743 (0.014)

0.643 (0.042)

0.828 (0.084)

0.719 (0.017)

0.721 (0.034)

KNN

0.538 (0.036)

0.704 (0.185)

0.599 (0.073)

0.545 (0.058)

0.537 (0.036)

0.704 (0.18)

0.599 (0.073)

0.546 (0.058)

RandomForest

0.710 (0.013)

0.839 (0.019)

0.769 (0.013)

0.810 (0.015)

0.711 (0.016)

0.841 (0.019)

0.770 (0.016)

0.811 (0.014)

XGBoost

0.702 (0.015)

0.806 (0.038)

0.750 (0.025)

0.794 (0.025)

0.697 (0.013)

0.815 (0.046)

0.751 (0.025)

0.797 (0.019)

SGD

0.617 (0.065)

0.649 (0.219)

0.604 (0.102)

0.605 (0.043)

0.597 (0.073)

0.617 (0.342)

0.534 (0.194)

0.568 (0.040)