Customer Event Rate Estimation Using Particle Filters

Estimating the rate at which events happen has been studied under various guises and in different settings. We are interested in the specific case of consumerinitiated events or transactions (credit/debit card transactions, mobile phone calls, online purchases, etc.), and the modeling of such behavior, in order to estimate the rate at which such transactions are made. In this paper, we detail a model of such events and a Bayesian approach, utilizing Sequential Monte Carlo technology, to online estimation of the event rate from event observations alone.

• Type: Conference
• Year: 2008
• Where: AUSDM
• Authors: Harsha Honnappa
• Keywords: AUSDM 2008, Credit/debit Card Transactions, Data Mining, Mobile Phone Calls, Sequential Monte Carlo