Benefit claims are a huge problem. The UK Government loses up to £7 billion
in mispaid Benefits [BA]. Detecting bogus claims amongst the immense
number of overall claims is a complex and time-consuming task. The
FEDAURA Project will develop
and evaluate new computerised methods for automated Benefit fraud detection.
The FEDAURA Project
will combine techniques from Statistics, Neural Networks and Machine
Learning to produce fraud detection technologies using large-scale Benefit
Claimant data sets. The FEDAURA Project
a baseline by evaluating currently used detection methods.
standards and produce an evaluation framework for fraud detection systems.
and evaluate a selected range of new techniques.
and evaluate the different methods for large-scale data sets.
the best methods via a single computer package and publications.
will be supported by the DTI Management of
Information Programme and the EPSRC (Grant No. GR/R55191/01) and will last for 36 months
from 1st November 2001. The
participating companies are:
BA internal Command paper, March 1999, A New Contract for Welfare: Safeguarding
Group, Department of Computer Science, University of York,
for Work & Pensions,