A new hybrid classification algorithm for customer churn prediction based on logistic regression and decision trees

Caigny, Arno De; Coussement, Kristof; Bock, Koen W. De

doi:10.1016/j.ejor.2018.02.009

Cited by 408 publications

(204 citation statements)

References 61 publications

Supporting

Mentioning

197

Contrasting

Unclassified

Order By: Relevance

“…Previous work [3,4] found that RF performs well in handling datasets with large class imbalance. However, the earlier studies used a large amount of training data, which improves the performance of RF to a certain extent.…”

Section: Discussionmentioning

confidence: 99%

“…The results indicated that the RF and GB perform very well and can cope with the class imbalance comparatively well. Caigny et al [4] proposed logit leaf model (LLM) to better classify imbalanced data. LLM was benchmarked against decision trees, LR, RF and logistic model tree (LMT) with regard to the predictive performance using fourteen churn datasets.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

The Effects of Class Imbalance and Training Data Size on Classifier Learning: An Empirical Study

Zheng

Jin

2020

SN COMPUT. SCI.

View full text Add to dashboard Cite

This study discusses the effects of class imbalance and training data size on the predictive performance of classifiers. An empirical study was performed on ten classifiers arising from seven categories, which are frequently employed and have been identified to be efficient. In addition, comprehensive hyperparameter tuning was done for every data to maximize the performance of each classifier. The results indicated that (1) naïve Bayes, logistic regression and logit leaf model are less susceptible to class imbalance while they have relatively poor predictive performance; (2) ensemble classifiers AdaBoost, XGBoost and parRF have a quite poorer stability in terms of class imbalance while they achieved superior predictive accuracies; (3) for all of the classifiers employed in this study, their accuracies decreased as soon as the class imbalance skew reached a certain point 0.10; note that although using datasets with balanced class distribution would be an ideal condition to maximize the performance of classifiers, if the skew is larger than 0.10, a comprehensive hyperparameter tuning may be able to eliminate the effect of class imbalance; (4) no one classifier shows to be robust to the change of training data size; (5) CART is the last choice among the ten classifiers.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

The Effects of Class Imbalance and Training Data Size on Classifier Learning: An Empirical Study

Zheng

Jin

2020

SN COMPUT. SCI.

View full text Add to dashboard Cite

show abstract

“…Huang and Kechadi (2013) presented a hybrid model-based learning system, which integrated the supervised and unsupervised techniques for predicting customer behaviour. A hybrid classification algorithm for customer churn prediction was proposed on the basis of logistic regression and decision trees (De Caigny, Coussement, & De Bock, 2018). Pendharkar (2009) applied two genetic algorithm-based NN models in the churn prediction context.…”

Section: Customer Churn Predictionmentioning

confidence: 99%

Churn prediction and social neighbour influences for different types of user groups in virtual worlds

et al. 2019

View full text Add to dashboard Cite

In virtual worlds (VWs), users have more VW games alternatives, whereas VW companies consequently suffer from high customer turnover rate and low customer loyalty. Therefore, building a churn prediction model to facilitate subsequent churn management and customer retention is important. The churn behaviours and the impact of social neighbour influences to customer churn may be different for different types of users. Accordingly, we segment users into stable, unstable, and solitary user groups according to their social contact behaviours in VWs. Novel segmentation‐based churn prediction approaches are proposed for churn prediction in VWs by building prediction models for each type of user groups and considering the effect of social neighbour influences for different user groups. The proposed approaches are evaluated by conducting experiments with a dataset collected from a VW platform. The experimental results show different churn prediction performances under different user groups. The segmentation‐based churn prediction approaches perform better than do general approaches without considering user groups. Moreover, the results also reveal that social neighbour influences have a positive impact on stable and unstable users. The proposed work contributes to investigating the social neighbour influences on churn prediction for different types of user groups in VWs.

show abstract

“…Recently, Umayaparvathi et al [22] wanted to find out the best among these CCP techniques by using decision tree, SVM and neural networks. SVM technique was found to be the best among them.…”

Section: Related Workmentioning

confidence: 99%

Improving the Reliability of Churn Predictions in Telecommunication Sector by Considering Customer Region

Zhou¹,

Boateng²,

Amoh³

et al. 2019

IJITCS

View full text Add to dashboard Cite

Prediction of customer churn has been one of the most interesting and challenging tasks facing most telecommunication companies recently. For the past decade, researchers together with practitioners have been working and designing models that can correctly make more accurate customer churn predictions (CCP). However, most of these models have less accuracy than expected which is hugely affecting the intended purpose. Consequently, most of these CCP models add little or nothing to the revenue growth of telecommunication industries. This work aims at improving the reliability of CCP in the telecommunication sector. To achieve this objective, a new attribute to be factored in CCP, known as the regional churn rate (RCR), is presented. Basically, RCR gives an idea about the rate of churning in a particular locality or region. Thus, a prediction model with a more accurate CCP using a support vector machine (SVM) classifier is proposed. The performance of the proposed model is critically evaluated using five metrics i.e. misclassification error, precision, recall, specificity and f-measure. At the same time, the performance of the proposed classifier (CCP with RCR) is compared with another SVM classifier which doesn't consider RCR (CCP without RCR). Results show that the proposed model which considers the RCR of a customer's location gives the highest accuracies for four performance metrics i.e. precision, recall, misclassification error and f-measure. Therefore, the proposed SVM-based CCP model gives a more clear indication as to whether a customer is a potential churner or not.

show abstract

A new hybrid classification algorithm for customer churn prediction based on logistic regression and decision trees

Cited by 408 publications

References 61 publications

The Effects of Class Imbalance and Training Data Size on Classifier Learning: An Empirical Study

The Effects of Class Imbalance and Training Data Size on Classifier Learning: An Empirical Study

Churn prediction and social neighbour influences for different types of user groups in virtual worlds

Improving the Reliability of Churn Predictions in Telecommunication Sector by Considering Customer Region

Contact Info

Product

Resources

About