- In business, “customer churn” is one of the most important metrics for a growing business to evaluate.
- Customer churn has many reasons and factors. Such reasons include quality and cost of services.
Ensemble learning
- Ensemle Learning is technique that craete multiple models and then combine them to produce improved results.
- Ensemble Learning usually produces more accurate solutions than a single model.
- Ensemble learning methods are applied to regression as well as classifacation.
- Voting Classifiers comes with multiple voting options such as hard and soft voting.
- Hard Voting uses multiple individual models to make its prediction.
-
Soft Voting relies on probabiltistic outcome values genereted by classification algorithms.
Comparison with other works
Further analysis
- All condidate models were evaluated on a public dataset in the telecom industry.
- MLP, Random Forest, CatBoost and XGBoost selected for an efficient stacking ensemble method.
- Created an ensemble model utilizing the soft voting approach.
- Proposed model showed the best accuracy of 88.2 % with balanced data.