Analytics for audit and business controls in corporate travel & entertainment

  • Authors:
  • Vijay Iyengar;Ioana Boier;Karen Kelley;Raymond Curatolo

  • Affiliations:
  • IBM Thomas J. Watson Research Center, Hawthorne, NY;IBM Thomas J. Watson Research Center, Hawthorne, NY;IBM Global Technology Services, Southbury, CT;IBM Global Technology Services, Southbury, CT

  • Venue:
  • AusDM '07 Proceedings of the sixth Australasian conference on Data mining and analytics - Volume 70
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Travel and Entertainment (T&E) expenses are under increasing scrutiny as one of the largest controllable indirect expenses in a firm. This involves internal audits and analysis by business controls personnel to identify fraud and misuse and to take appropriate corrective actions. We have developed a set of statistical models to identify suspicious behavior for further investigation. Our Behavioral Shift Models (BSM) leverage domain knowledge in the form of simple, generic templates that represent classes of fraud and abuse. The emphasis is on robustly detecting repeated, out-of-the-norm behaviors as opposed to single instance occurrences. In this paper, we describe the application of these models and characterize their detection capabilities empirically. We also present validated results and insights generated by our approach when applied to production data from multiple firms for several T&E scenarios.