Feature-selection-based dynamic transfer ensemble model for customer churn prediction

详细信息    查看全文

• 作者：Jin Xiao (1)
  Yi Xiao (2)
  Anqiang Huang (3)
  Dunhu Liu (4)
  Shouyang Wang (5)
  
  1. Business School ; Sichuan University ; Chengdu ; 610064 ; China
  2. School of Information Management ; Central China Normal University ; Wuhan ; 430079 ; China
  3. School of Economics and Management ; Beihang University ; Beijing ; 100083 ; China
  4. Management Faculty ; Chengdu University of Information Technology ; Chengdu ; 610103 ; China
  5. Academy of Mathematics and Systems Science ; Chinese Academy of Sciences ; Beijing ; 100190 ; China
  
• 关键词：Customer churn prediction ; Transfer ensemble model ; Feature selection ; GMDH ; type neural network ; Transfer learning
• 刊名：Knowledge and Information Systems
• 出版年：2015
• 出版时间：April 2015
• 年：2015
• 卷：43
• 期：1
• 页码：29-51
• 全文大小：756 KB
• 参考文献：1. Dych茅, J (2001) The CRM handbook: a business guide to customer relationship management. Addison-Wesley, Reading
  2. Bhattacharya, CB (1998) When customers are members: customer retention in paid membership contexts. J Acad Market Sci 26: pp. 31-44 2070398261004" target="_blank" title="It opens in new window">CrossRef
  3. Neslin, SA, Gupta, S, Kamakura, W, Lu, JX, Mason, CH (2006) Detection defection: measuring and understanding the predictive accuracy of customer churn models. J Market Res 43: pp. 204-211 2.204" target="_blank" title="It opens in new window">CrossRef
  4. Au, W, Chan, KCC, Yao, X (2004) A novel evolutionary data mining algorithm with applications to churn prediction. IEEE T Evol Comput 7: pp. 532-545
  5. Kisioglu, P, Topcu, YI (2011) Applying Bayesian belief network approach to customer churn analysis: a case study on the telecom industry of Turkey. Expert Syst Appl 38: pp. 7151-7157 2010.12.045" target="_blank" title="It opens in new window">CrossRef
  6. Pendharkar, PC (2005) A threshold-varying artificial neural network approach for classification and its application to bankruptcy prediction problem. Comput Oper Res 32: pp. 2561-2582 2004.06.023" target="_blank" title="It opens in new window">CrossRef
  7. Wei, CP, Chiu, IT (2002) Turning telecommunications call details to churn prediction: a data mining approach. Expert Syst Appl 23: pp. 103-112 2)00030-1" target="_blank" title="It opens in new window">CrossRef
  8. Zhao, Y, Li, B, Li, X, Liu, W, Ren, S Customer churn prediction using improved one-class support vector machine. In: Li, X, Wang, S, Dong, ZY eds. (2005) ADMA 2005, LNAI 3584. Springer, Berlin, pp. 300-306
  9. Wang, BX, Japkowicz, N (2010) Boosting support vector machines for imbalanced data sets. Knowl Inf Syst 25: pp. 1-20 CrossRef
  10. Yang, Q, Wu, X (2006) 10 challenging problems in data mining research. Int J Inf Tech Decis 5: pp. 597-604 2/S0219622006002258" target="_blank" title="It opens in new window">CrossRef
  11. Verbeke, W, Martens, D, Mues, C, Baesens, B (2011) Building comprehensible customer churn prediction models with advanced rule induction techniques. Expert Syst Appl 38: pp. 2354-2364 2010.08.023" target="_blank" title="It opens in new window">CrossRef
  12. Xia, G, Jin, W (2008) Model of customer churn prediction on support vector machine. Syst Eng Theor Pract 28: pp. 71-77 CrossRef
  13. Lemmens, A, Croux, C (2006) Bagging and boosting classification trees to predict churn. J Market Res 43: pp. 276-286 2.276" target="_blank" title="It opens in new window">CrossRef
  14. Glady, N, Baesens, B, Croux, C (2009) Modeling churn using customer lifetime value. Eur J Oper Res 197: pp. 402-411 2008.06.027" target="_blank" title="It opens in new window">CrossRef
  15. Vapnik, V (1998) Statistical learning theory. Wiley, New York
  16. Pan, SJ, Yang, Q (2010) A survey on transfer learning. IEEE T Knowl Data En 22: pp. 1345-1359 2009.191" target="_blank" title="It opens in new window">CrossRef
  17. Kittler, J, Hatef, M, Duin, RPW, Matas, J (1998) On combining classifiers. IEEE T Pattern Anal 20: pp. 226-239 CrossRef
  18. Amanifard, N, Nariman-Zadeh, N, Borji, M, Khalkhali, A, Habibdoust, A (2008) Modelling and Pareto optimization of heat transfer and flow coefficients in microchannels using GMDH type neural networks and genetic algorithms. Energ Convers Manag 49: pp. 311-325 2007.06.002" target="_blank" title="It opens in new window">CrossRef
  19. Ivakhnenko, AG (1976) The group method of data handling in prediction problems. Soviet Autom Contr 9: pp. 21-30
  20. Ranawana, R, Palade, V (2006) Multi-classifier systems: review and a roadmap for developers. Int J Hybr Intell Syst 3: pp. 35-61
  21. Hansen, LK, Salamon, P (1990) Neural network ensembles. IEEE T Pattern Anal 12: pp. 993-1001 CrossRef
  22. Woods, K, Kegelmeyer, WP, Bowyer, K (1997) Combination of multiple classifiers using local accuracy estimates. IEEE T Pattern Anal 19: pp. 405-410 27" target="_blank" title="It opens in new window">CrossRef
  23. Kuncheva, L, Whitaker, C (2003) Measures of diversity in classifier ensembles and their relationship with the ensemble accuracy. Mach Learn 51: pp. 181-207 23/A:1022859003006" target="_blank" title="It opens in new window">CrossRef
  24. Breiman, L (1996) Bagging predictors. Mach Learn 24: pp. 123-140
  25. Ho, TK (1998) The random space method for constructing decision forests. IEEE T Pattern Anal 20: pp. 832-844 CrossRef
  26. Zhu, X, Wu, X, Yang, Y (2006) Effective classification of noisy data streams with attribute-oriented dynamic classifier selection. Knowl Inf Syst 9: pp. 339-363 212-y" target="_blank" title="It opens in new window">CrossRef
  27. Ko AHR, Sabourin R, Britto AS Jr (2008) From dynamic classifier selection to dynamic ensemble selection. Pattern Recogn 41(5):1718鈥?731
  28. Bi W, Shi Y, Lan Z (2009) Transferred feature selection. In: Proceedings of IEEE international conference on data mining workshops, pp 416鈥?21
  29. Kamishima T, Hamasaki M, Akaho S (2009) TrBagg: a simple transfer learning method and its application to personalization in collaborative tagging. In: Proceedings of ninth IEEE international conference on data mining, Miami, FL, USA, pp 219鈥?28
  30. Dai W, Yang Q, Xue GR, Yu Y (2007) Boosting for transfer learning. In: Proceedings of the 24th international conference on machine learning, pp 193鈥?00
  31. Freund, Y, Schapire, RE (1997) A decision-theoretic generalization of on-line learning and an application to boosting. J Comput Syst Sci 55: pp. 119-139 CrossRef
  32. Mueller JA, Lemke F (2000) Self-organising data mining: an intelligent approach to extract knowledge from data. Libri
  33. Abdel-Aal, RE, Elhadidy, MA, Shaahid, SM (2008) Modeling and forecasting the mean hourly wind speed time series using GMDH-based abductive networks. Renew Energ 34: pp. 1686-1699 2009.01.001" target="_blank" title="It opens in new window">CrossRef
  34. Puig, V, Witczak, M, Nejjari, F, Quevedo, J, Korbicz, J (2007) A GMDH neural network-based approach to passive robust fault detection using a constraint satisfaction backward test. Eng Appl Artif Intell 20: pp. 886-897 2006.12.005" target="_blank" title="It opens in new window">CrossRef
  35. Xiao, J, He, CZ, Jiang, XY, Liu, DH (2010) A dynamic classifier ensemble selection approach for noise data. Inform Sci 180: pp. 3402-3421 2010.05.021" target="_blank" title="It opens in new window">CrossRef
  36. Xiao, J, Xie, L, He, CZ, Jiang, XY (2012) Dynamic classifier ensemble model for customer classification with imbalanced class distribution. Expert Syst Appl 39: pp. 3668-3675 2011.09.059" target="_blank" title="It opens in new window">CrossRef
  37. He, CZ (2005) Self-organising data mining and economic forecasting. Science Publish, Beijing
  38. Merz C, Murphy P (1995) UCI repository of machine learning databases. http://www.ics.uci.edu/~mlearn/MLRepository.html
  39. Friedman JH (2003) On multivariate goodness-of-fit and two-sample testing. In: Proceedings of Phystat 2003. SLAC, Stanford, CA, pp 1鈥?
  40. Cortes, C, Vapnik, V (1995) Support vector networks. Mach Learn 20: pp. 273-297
  41. Tsymbal, A, Puuronen, S, Patterson, DW (2003) Ensemble feature selection with the simple Bayesian classification. Inform Fusion 4: pp. 87-100 2535(03)00004-6" target="_blank" title="It opens in new window">CrossRef
  42. Doumpos, M, Zopounidis, C (2004) A multicriteria classification approach based on pairwise comparisons. Eur J Oper Res 158: pp. 378-389 2003.06.011" target="_blank" title="It opens in new window">CrossRef
  43. Poel, D, Buckinx, W (2005) Predicting online-purchasing behaviour. Eur J Oper Res 166: pp. 557-575 2004.04.022" target="_blank" title="It opens in new window">CrossRef
  44. McNemar, Q (1947) Note on the sampling error of differences between correlated proportions and percentages. Psychometrica 12: pp. 153-157 2295996" target="_blank" title="It opens in new window">CrossRef
  
• 刊物类别：Computer Science
• 刊物主题：Information Systems and Communication Service
  Business Information Systems
  
• 出版者：Springer London
• ISSN：0219-3116

文摘

Customer churn prediction is one of the key steps to maximize the value of customers for an enterprise. It is difficult to get satisfactory prediction effect by traditional models constructed on the assumption that the training and test data are subject to the same distribution, because the customers usually come from different districts and may be subject to different distributions in reality. This study proposes a feature-selection-based dynamic transfer ensemble (FSDTE) model that aims to introduce transfer learning theory for utilizing the customer data in both the target and related source domains. The model mainly conducts a two-layer feature selection. In the first layer, an initial feature subset is selected by GMDH-type neural network only in the target domain. In the second layer, several appropriate patterns from the source domain to target training set are selected, and some features with higher mutual information between them and the class variable are combined with the initial subset to construct a new feature subset. The selection in the second layer is repeated several times to generate a series of new feature subsets, and then, we train a base classifier in each one. Finally, a best base classifier is selected dynamically for each test pattern. The experimental results in two customer churn prediction datasets show that FSDTE can achieve better performance compared with the traditional churn prediction strategies, as well as three existing transfer learning strategies.