wir bieten...
Dekobild im Seitenkopf ISMLL
 
Courses in winter term 2007/2008 / Master Seminar on Fraud Detection
Abstract

Time: Wed. 16:15 - 18:00
Location: C 213 Spl
Begin: Wed., October 24th, 2007

The extent of fraud is difficult to quantify, but we are sure, it is lot of millions of euro, and this amount is growing rapidly. For example only in the UK the estimated amount of credit card fraud has been growing from 122 millions of pound (1997) to almost 300 millions during three years. (Bolton und Hand: Statistical Fraud Detection: A Review).

In this seminar we discuss different machine learning techniques and concepts for automatical fraud detection.

Contact: Prof. Dr. Dr. Lars Schmidt-Thieme und Krisztian Buza

Topics:

  1. -- Introduction --
  2. Fraud Detection in Recommender Systems
  3. Credit Card Fraud Detection with Bayesian and Neural Networks
  4. Fraud Detection based on Data Signature
  5. Fraud Detection in Sequential Data
  6. Case Based Systems for Fraud Detection
  7. Relational Classification for Fraud Detection
  8. Boosting Naive Bayes for Fraud Detection
  9. Improvement of Internal Control Systems
  10. Regression for Prediction of Bankruptcy
  11. Hybrid Systems (Deceison Trees and SVM)

Die Themen werden per E-Mail vergeben ( ) .