Additional Reading: ICS 278: Data Mining

Spring 2006

Below I will provide links to papers and Web pages that complement the material presented in class. Additional links will be added as the quarter progresses.

• · General Background:
• How much information is in the world? Michael Lesk, 1998. A little outdated in terms of the specific numbers, but nonetheless an interesting article.
  

• · Logistic Regression
• Fast robust logistic regression for large sparse data sets with binary outputs, Komarek and Moore, 2003
• Making logistic regression a core data mining tool, Komarek and Moore, 2005.  
  
  

• · Fraud detection
• Overview Web page on fraud detection methods from AAAI
  

• · Topic Models for Text Data:
• M. Steyvers and T. Griffiths, Probabilistic topic models, In T. Landauer, D McNamara, S. Dennis, and W. Kintsch (eds), Latent Semantic Analysis: A Road to Meaning. Laurence Erlbaum, in press.
• D. Blei, M. Jordan, A. Ng, Latent Dirichlet allocation, JMLR, 2003
• M. Rosen-Zvi, T. Griffiths, M. Steyvers, and P. Smyth, The author-topic model for authors and documents, Proceedings of the 20^th Conference on Uncertainty in AI, 2004.
• Web page for the author-topic model, including an online JAVA browser for the NIPS and CiteSeer data sets.
  
• Papers from the Helsinki Institute of Information Technology ALVIS project for topic models and Web search
  
  
• · AskMSR question-answering system:
• S. Dumais et al., Web question answering: Is more always better? Proceedings of ACM SIGIR 2002
• Popular technology article in MIT Technology Review on Web search in general, 2004.
  
  
• · Credit Scoring:
• D. J. Hand and W. E. Henley, Statistical Classification Methods in Consumer Credit Scoring: A Review? J. Royal Statistical Society A, 1997.

• Links to articles on naïve Bayes and spam email:

· Popfile: a popular open-source email filtering program using naïve Bayes classification

· Paul Graham’s widely-cited article, A plan for spam

· Upcoming first academic conference on email and anti-spam

· Technical paper on spam email classification (comparing naïve Bayes with memory-based classification methods)

· SpamBayes: another open source filtering program using Bayes rule

· Mozilla page on spam email (again using naïve Bayes classification)

· SpamAssassin: a rule-based spam filtering approach