Profit-based churn prediction based on Minimax Probability Machines

Maldonado, Sebastián; López, Julio; Vairetti, Carla

doi:10.1016/j.ejor.2019.12.007

Cited by 39 publications

(11 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…e model outperformed state-of-the-art based on the article's results. Maldonado et al [16] provide a new churn prediction method that drives profit. e method is a developed version of the minimax probability machine algorithm for the classification of churns.…”

Section: Literature Review Of Churn Prediction Methodsmentioning

confidence: 99%

“…e study of principal components is one of the outcomes of linear algebra mathematics because the nonparametric and straightforward method extracts relevant information from confusing sets. e transformation of the T can be obtained by minimizing the least-squares error, assuming that the CCPBI-TAMO, CPIO-FS Telecom Precision, recall, accuracy, F-Score, ROC [41] Xgboost, AdaBoost, catboost, decision trees, SVM, KNN Telecom Accuracy, AUC, precision, recall, F-Measures [37] Deep feed-forward networks Subscription companies Accuracy [38] Deep ANN, machine learning algorithms Telecom Accuracy, precision, recall, F1-score, and AUC [12] Neural network with bagging Telecom Accuracy, precision, recall, F-score, kappa, absolute error, relative error, and classi cation error [10] Transfer learning of ensemble Telecom Area under curve of ROC (AUC) and complexity [11] Ensemble algorithm Telecom Area under curve of ROC (AUC) [12] Begging and neural network Telecom Accuracy and precision of classi cation [42] Arti cial neural network (ANN) and self-organized map (SOM) Telecom Accuracy, recall, F-score, and precision [15] Pro t tree Telecom Accuracy, cost, and pro t [16] Minimax probability machines Telecom AUC and EMPC [17] similarity forests Telecom AUC, and tenlift AUPR [21] Temporal point processes (TPP) and recurrent neural networks (RNN) Telecom MAE and MRE [22] Cross-company just-in-time approach Telecom Accuracy, Kappa, and Recall [25] Multiobjective and colony optimization Telecom AUC [27] graph theory Telecom Top decile lift [31] Boosted…”

Section: Principal Component Analysis (Pca)mentioning

confidence: 99%

See 1 more Smart Citation

Model Optimization Analysis of Customer Churn Prediction Using Machine Learning Algorithms with Focus on Feature Reductions

Mirabdolbaghi

Amiri

2022

Discrete Dynamics in Nature and Society

View full text Add to dashboard Cite

Currently, Customers are struggling to retain their business in today’s competitive markets. Thus, the issue of customer churn becomes a significant challenge for the industries. In order to achieve this, it is vital to have an efficient churn prediction system. In this paper, we discuss methods for reducing features using PCA, Autoencoders, LDA, T-SNE, and Xgboost. In this paper, a model for predicting light GBM churn is proposed. The model consists of five steps. The first step is to preprocess the data so that missing and corrupt values can be handled and the data can be scaled. Secondly, implementing a comprehensive feature reduction system based on popular algorithms reduces the features and selects the most suitable one. In the third step, light GBM’s hyperparameter is tuned using Bayesian hyperparameter optimization and genetic optimization algorithms. Lastly, interpreting the model and evaluating the impact of the features on model outputs by using the SHAP method, and finally ranking the churners by customer lifetime value. Aside from evaluating and choosing the best feature reduction methods, the proposed method is also evaluated using four famous datasets. It outperforms other ensemble and ML algorithms like AdaBoost, SVM, and decision tree on over seven evaluation metrics: accuracy, area under the curve (AUC), Kappa, Mathews correlation coefficient (MCC), Brier score, F1 score, and EMPC. In light of the evaluation metrics, our model shows a significant improvement in handling imbalanced datasets in churn prediction. Finally, in this paper, interpretability and how the features affect the model’s output are presented by the SHAP method. Then CLV ranking is suggested for better decision-making.

show abstract

Section: Literature Review Of Churn Prediction Methodsmentioning

confidence: 99%

Section: Principal Component Analysis (Pca)mentioning

confidence: 99%

Model Optimization Analysis of Customer Churn Prediction Using Machine Learning Algorithms with Focus on Feature Reductions

Mirabdolbaghi

Amiri

2022

Discrete Dynamics in Nature and Society

View full text Add to dashboard Cite

show abstract

“…Traditional statistical modelling in churn prediction tends to use models like logistic regression, survival models, neural networks and self-organising maps (Vafeiadis et al , 2015; Verbeek, 2015; Klepac, 2014), all of which require existing customer data. One problem with using existing company data for churn prediction is that sometimes data are not time-stamped and, therefore, are not usable for tracking retention campaigns (Maldonado et al , 2020). In addition, data must be available, and it is extremely unlikely that competitor or industry-wide company data will be made available for churn analysis.…”

Section: Literature Reviewmentioning

confidence: 99%

What social media sentiment tells us about why customers churn

Lappeman

Franco

Warner

et al. 2022

JCM

View full text Add to dashboard Cite

Purpose This study aims to investigate the factors that influence South African customers to potentially switch from one bank to another. Instead of using established models and survey techniques, the research measured social media sentiment to measure threats to switch. Design/methodology/approach The research involved a 12-month analysis of social media sentiment, specifically customer threats to switch banks (churn). These threats were then analysed for co-occurring themes to provide data on the reasons customers were making these threats. The study used over 1.7 million social media posts and focused on all five major South African retail banks (essentially the entire sector). Findings This study concluded that seven factors are most significant in understanding the underlying causes of churn. These are turnaround time, accusations of unethical behaviour, billing or payments, telephonic interactions, branches or stores, fraud or scams and unresponsiveness. Originality/value This study is unique in its measurement of unsolicited social media sentiment as opposed to most churn-related research that uses survey- or customer-data-based methods. In addition, this study observed the sentiment of customers from all major retail banks across 12 months. To date, no studies on retail bank churn theory have provided such an extensive perspective. The findings contribute to Susan Keaveney’s churn theory and provide a new measurement of switching threat through social media sentiment analysis.

show abstract

“…Later, Al-Mashraie et al [22] compared multiple data analytic approaches for the prediction of telecommunication customer switching behaviour in the United States to demonstrate that the support vector machine (SVM) model outperformed the LR, random forest (RF), and decision tree (DT) models. Maldonado et al [23] employed a profit-driven data analytic approach to improve the economic effectiveness of predictions of customer churn rates. In addition to the above default prediction domains, the existing literature also reported on financial statement fraud detection [24], [25].…”

Section: Literature Reviewmentioning

confidence: 99%

Tax Default Prediction Using Feature Transformation-Based Machine Learning

et al. 2021

View full text Add to dashboard Cite

This study proposes to address the economic significance of unpaid taxes by using an automatic system for predicting a tax default. Too little attention has been paid to tax default prediction in the past. Moreover, existing approaches tend to apply conventional statistical methods rather than advanced data analytic approaches, including state-of-the-art machine learning methods. Therefore, existing studies cannot effectively detect tax default information in real-world financial data because they fail to take into account the appropriate data transformations and nonlinear relationships between early-warning financial indicators and tax default behavior. To overcome these problems, this study applies diverse feature transformation techniques and state-of-the-art machine learning approaches. The proposed prediction system is validated by using a dataset showing tax defaults and non-defaults at Finnish limited liability firms. Our findings provide evidence for a major role of feature transformation, such as logarithmic and square-root transformation, in improving the performance of tax default prediction. We also show that extreme gradient boosting and the systematically developed forest of multiple decision trees outperform other machine learning methods in terms of accuracy and other classification performance measures. We show that the equity ratio, liquidity ratio, and debt-to-sales ratio are the most important indicators of tax defaults for 1-year-ahead predictions. Therefore, this study highlights the essential role of well-designed tax default prediction systems, which require a combination of feature transformation and machine learning methods. The effective implementation of an automatic tax default prediction system has important implications for tax administration and can assist administrators in achieving feasible government expenditure allocations and revenue expansions.INDEX TERMS Default prediction, corporate tax, machine learning, feature transformation.

show abstract

Profit-based churn prediction based on Minimax Probability Machines

Cited by 39 publications

References 31 publications

Model Optimization Analysis of Customer Churn Prediction Using Machine Learning Algorithms with Focus on Feature Reductions

Model Optimization Analysis of Customer Churn Prediction Using Machine Learning Algorithms with Focus on Feature Reductions

What social media sentiment tells us about why customers churn

Tax Default Prediction Using Feature Transformation-Based Machine Learning

Contact Info

Product

Resources

About