Abstract
Some predictive models for customer value management might benefit from information about certain changes in individual-consumer behaviour. We take changepoint methods as the first step in producing a model-input feature for this purpose. An unusual feature in the application of changepoint methods to consumer data is there are as many streams of data as there are customers. This property is used to help decide whether an individual has changed their behaviour by ordering likelihood-ratio statistics from the changepoint models. Following a review of changepoint methods, the approach is demonstrated on cash machine transactions. Models for the amount, location and time of transaction are used and accounts exhibiting large evidence of change are examined in detail. For the data set used the approach performs sensibly. The worth of likelihood-ratio statistics to rank evidence for change is considered more generally through some of the literature.
References
Barnard GA (1959). Control charts and stochastic processes . J Roy Stat Soc B 21: 239–271.
Bélisle P, Joseph L, Macgibbon B, Wolfson DB and du Berger R (1998). Change-point analysis of neuron spike train data . Biometrics 54: 113–123.
Brentnall AR, Crowder MJ and Hand DJ (2008). A statistical model for the temporal pattern of individual automated teller machine withdrawals . Appl Stat 57: 43–59.
Brentnall AR, Crowder MJ and Hand DJ (2010). Predicting the amount individuals withdraw at cash machines using a multinomial random-effects model. Statistical Modelling, in press.
Carlin BP, Gelfand AE and Smith AFM (1992). Hierarchical Bayesian analysis of changepoint problems . Appl Stat 41: 389–405.
Carlstein E (1988). Nonparametric change-point estimation . Ann Stat 16: 188–197.
Chen J and Gupta AK (2000). Parametric Statistical Change Point Analysis . Birkhäuser Boston: New York.
Chen J and Gupta AK (2001). On change point detection and estimation . Commun Stat: Simul C 30: 665–697.
Csörgö M and Horváth L (1997). Limit Theorems in Change-Point Analysis . John Wiley & Sons: Chicester, England.
Fader P, Hardie B and Huang C (2004). A dynamic changepoint model for new product sales forecasting . Market Sci 23: 50–65.
Gustafsson F (2000). Adaptive Filtering and Change Detection . John Wiley & Sons: Chicester, England.
Hand DJ, Brentnall AR and Crowder MJ (2008). Credit scoring: A future beyond empirical models . J Financ Transform 23: 121–128.
Hinkley DV (1970). Inference about the change-point in a sequence of random variables . Biometrika 57: 1–17.
Hinkley DV (1971). Inference about the change-point from cumulative sum tests . Biometrika 58: 509–523.
James B, James KL and Siegmund D (1987). Tests for a change-point . Biometrika 74: 71–83.
Joseph L, Wolfson DB, du Berger R and Lyle RM (1997). Analysis of panel data with change-points . Stat Sinica 7: 687–703.
Kiuchi AS, Hartigan JA, Holford TR, Rubinstein P and Stevens CE (1995). Change points in the series of T4 counts prior to AIDS . Biometrics 51: 236–248.
Legal Services Research Centre (2007). A helping hand: The impact of debt advice on people's lives. Technical report, Legal Services Research Centre, http://www.lsrc.org.uk/publications/Impact.pdf, accessed 7 December 2007.
Mira A and Petrone S (1996). Bayesian hierarchical non-parametric inference for change-point problems . In: Bernardo JM, Berger JO, Dawid AP and Smith AFM (eds). Bayesian Statistics 5. Oxford University Press: USA, pp. 693–703.
Moustakides GV (1986). Optimal stopping times for detecting changes in distributions . Ann Stat 14: 1379–1387.
Page ES (1954). Continuous inspection schemes . Biometrika 41: 100–115.
Pettitt AN (1979). A non-parametric approach to the change-point problem . Appl Stat 28: 126–135.
Romano JP and Wolf M (2005). Stepwise multiple testing as formalized data snooping . Econometrica 73: 1237–1282.
Sen A and Srivastava MS (1975). On tests for detecting change in mean . Ann Statist 3: 98–108.
Siegmund D (1986). Boundary crossing probabilities and statistical applications . Ann Statist 14: 361–404.
Smith AFM (1975). A Bayesian approach to inference about a change-point in a sequence of random variables . Biometrika 62: 407–416.
Thomas LC, Mues C and Matuszyk A (2007). Modelling LGD for unsecured personal loans: Decision tree approach. University of Southampton Discussion Paper Series: Centre for Operational Research, Management Sciences and Information Systems, CORMSIS-07-07, http://eprints.soton.ac.uk/58330/, accessed 26 June 2009.
Worsley KJ (1983). The power of likelihood ratio and cumulative sum tests for a change in a binomial probability . Biometrika 70: 455–464.
Yi G, Coleman S and Ren Q (2006). CUSUM method in predicting regime shifts and its performance in different stock markets allowing for transaction fees . J Appl Stat 33: 647–661.
Acknowledgements
This work is supported by EPSRC grant number EP/D505380/1 awarded to the Quantitative Financial Risk Management Centre at Imperial College London.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Brentnall, A., Crowder, M. & Hand, D. Likelihood-ratio changepoint features for consumer-behaviour models. J Oper Res Soc 61, 462–472 (2010). https://doi.org/10.1057/jors.2009.160
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1057/jors.2009.160