Data mining is currently becoming an increasingly hot research field, but a large gap still remains between the research of data mining and its application in real-world business. As one of the largest data users in Australia, Centrelink has huge volume of data in data warehouse and tapes. Based on the available data, Centrelink is seeking to find underlying patterns to be able to intervene earlier to prevent or minimize debt. To discover the debtor patterns of Centrelink customers and bridge the gap between data mining research and application, we have done a project on improving income reporting to discover the patterns of those customers who were or are in debt to Centrelink. Two data models were built respectively for demographic data and activity data, and decision tree and sequence mining were used respectively to discover demographic patterns and activity sequence patterns of debtors. The project produced some potentially interesting results, and paved the way for more data mining applications in Centrelink in near future.
|Cite as: Zhao, Y., Cao, L., Morrow, Y., Ou, Y., Ni, J. and Zhang, C. (2006). Discovering Debtor Patterns of Centrelink Customers. In Proc. Fifth Australasian Data Mining Conference (AusDM2006), Sydney, Australia. CRPIT, 61. Peter, C., Kennedy, P. J., Li, J., Simoff, S. J. and Williams, G. J., Eds. ACS. 135-144. |