Modelling telephone call center arrivals as a nonhomogenous poisson process with cyclic rate

Abstract

This project demonstrates that telephone call center call-arrival data can be effectively modelled as anonhomogenous Poisson process with cyclic rate. The data was from an Israeli bank telephone call center collected through the year 1999 and made freely accessible online. The theory underlying the assumptions of a Poisson process has been presented where it has been shown that time-sampling a Poisson process results in a nonhomogenous Poisson process and the method of maximum likelihood presented and applied to estimate the parameters of an exponential-polynomial-trigonometric rate function (EPTF). An analysis of this data confirmed that a nonhomogenous Poisson process with an EPTF-type rate was a good fit with a confidence of 90%.