KNOWLEDGE MINING:
  Knowledge mining with neural networks and support vector machines for the development of customer scoring models

 

Financing: Internal Funding KU Leuven (KU Leuven)

Project reference Nr.: OT/03/12
Start: 2003-10-01
End: 2007-09-30

Description:
Recent developments of powerful classification techniques, combined with the availability of large client-credit databases, create new possibilities for customer-scoring problems, e.g. in the financial or marketing domains. We want to use neural networks and support vector machines for both domains in an economical relevant framework. Topics to be studied are efficient modelling of misclassification costs, measurement and comparison of model behaviour and development of facilities for input selection. The main goal is the development of user-friendly and powerful decision systems to improve the efficiency of the economic process.
 

SMC people involved in the project: