tc.bib

@INPROCEEDINGS{apte94towards,
  AUTHOR = {Chidanand Apte and Fred Damerau and Sholom M. Weiss},
  TITLE = {Towards Language Independent Automated Learning of Text Categorisation Models},
  BOOKTITLE = {Research and Development in Information Retrieval},
  PAGES = {23--30},
  YEAR = {1994},
  ANNOTE = {The authors present the results of rule-based two-class text categorization for English and German corpuses.}
}

@ARTICLE{apte94automated,
  AUTHOR = {Chidanand Apte and Fred Damerau and Sholom M. Weiss},
  TITLE = {Automated Learning of Decision Rules for Text Categorization},
  JOURNAL = {Information Systems},
  VOLUME = {12},
  NUMBER = {3},
  PAGES = {233--251},
  YEAR = {1994},
  URL = {http://citeseer.nj.nec.com/apte94automated.html},
  ANNOTE = {The content is essentially same as above. Rule-based method SWAP-1 is presented more thoroughly. The paper contains also a concise introduction to text categorization. }
}

@ARTICLE{brewlow97simplifying,
  AUTHOR = {L. Brewlow and D. Aha},
  TITLE = {Simplifying decision trees: a survey},
  JOURNAL = {Knowledge Engineering Review},
  VOLUME = {12},
  NUMBER = {1},
  PAGES = {1--40},
  YEAR = {1997},
  URL = {http://citeseer.nj.nec.com/14866.html},
  ANNOTE = {Authors review several methods of simplifying decision trees and discuss relating problems. Five categories: control tree size, expand the set of teset, modified test research, database restrictions, and alternative data structures.}
}

@ARTICLE{chen95machine,
  AUTHOR = {Hsinchun Chen},
  TITLE = {Machine Learning for Information Retrieval: Neural Networks, Symbolic Learning, and Genetic Algorithms},
  JOURNAL = {Journal of the American Society of Information Science},
  VOLUME = {46},
  NUMBER = {3},
  PAGES = {194--216},
  YEAR = {1995},
  URL = {http://ai.bpa.arizona.edu/papers/PS/mlir93.ps},
  ANNOTE = {Chen presents Hopfield networks, ID3/ID5 and genetic algorithms in the context of information retrieval. There is a fairly elaborate introduction to information retrieval, as well as the methods mentioned above. Extensive bibliography.}
}

@INPROCEEDINGS{germany98inductive,
  AUTHOR = {Susan T. Dumais and
               John Platt and
               David Hecherman and
               Mehran Sahami},
  TITLE = {Inductive learning algorithms and representations for text categorization},
  BOOKTITLE = {Proc. 7th International Conference on Information and Knowledge Management CIKM},
  PAGES = {148--155},
  YEAR = {1998},
  URL = {robotics.stanford.edu/users/sahami/papers-dir/cikm98.pdf},
  ANNOTE = {Naive Bayes, Rocchio, Bayes Nets, SVM and decision trees are experimented. }
}

@INPROCEEDINGS{lam98using,
  AUTHOR = {Wai Lam and Chao Y. Ho},
  TITLE = {Using a generalized instance set for automatic text categorization},
  BOOKTITLE = {Proc. {ACM SIGIR}-98},
  PAGES = {81--89},
  YEAR = {1998},
  ANNOTE = {GIS is an improvement to kNN. Nearest Neighbour in typically lazy and employs all of the native space while classifying, whereas GIS combines the strengths of kNN and linear classifiers through generalizing. Implementation: ExpNet.}
}

@INPROCEEDINGS{langley94oblivious,
  AUTHOR = {Langley and Stephanie Sage},
  TITLE = {Oblivious Decision Trees and Abstract Cases},
  BOOKTITLE = {Proc. AAAI-94 Workshop on case-based reasoning},
  PAGES = {113--117},
  YEAR = {1994},
  URL = {http://www.isle.org/~langley/papers/oblivion.cbr94.ps.gz},
  ANNOTE = {Oblivion is a decision tree capable of dealing with irrelevant data. Tests with Oblivion, kNN and C4.5}
}

@INPROCEEDINGS{lewis94comparison,
  AUTHOR = {David D. Lewis and Marc Ringuette},
  TITLE = {A comparison of two learning algorithms for text categorization},
  BOOKTITLE = {Proc. Symposium on Document Analysis and Information Retrieval {SDAIR}-94 },
  PAGES = {81--93},
  YEAR = {1994},
  URL = {citeseer.nj.nec.com/lewis94comparison.html},
  ANNOTE = {Authors compare a Bayesian classifier with decision tree.}
}

@INPROCEEDINGS{lewis98naive,
  AUTHOR = {David Lewis},
  TITLE = {Naive Bayes at forty: The independence assumption in information retrieval},
  BOOKTITLE = {Proc. 10th European Conference on Machine Learning ECML-98},
  PAGES = {4--15},
  YEAR = {1998},
  URL = {http://www.research.att.com/~lewis/papers/lewis98b.ps},
  ANNOTE = {First Lewis shortly presents idea of Naive Bayes. Then he compares the binary independence model (Multivariate Bernoulli) and then Multinomial Model.}
}

@INPROCEEDINGS{mccallum98comparison,
  AUTHOR = {Andrew McCallum and K. Nigam},
  TITLE = {A comparison of event models for Naive Bayes text classification},
  BOOKTITLE = {Proc. AAAI-98 Workshop on Learning for Text Categorization},
  YEAR = {1998},
  URL = {citeseer.nj.nec.com/mccallum98comparison.html},
  ANNOTE = {The paper compares the multivariate Bernoulli with multinomial event model. These both make the naive independence assumption, but the latter captures the word frequency whereas the former has a strict binary representation.}
}

@INPROCEEDINGS{sahami96applying,
  AUTHOR = {Mehran Sahami and Marti Hearst and Eric Saund},
  TITLE = {Applying the Multiple Cause Mixture Model to Text Categorization},
  BOOKTITLE = {Proc. 13th International Conference on Machine Learning ICML96},
  PAGES = {435--443},
  YEAR = {1996},
  URL = {citeseer.nj.nec.com/sahami96applying.html},
  ANNOTE = {The paper introduces a method that can be used for both supervised and unsupervised learning. The performance is not quite appealing.}
}

@INPROCEEDINGS{yang97comparative,
  AUTHOR = {Yiming Yang and Jan O. Pedersen},
  TITLE = {A comparative study on feature selection in text categorization},
  BOOKTITLE = {Proc. 14th International Conference on Machine Learning},
  PAGES = {412--420},
  YEAR = {1997},
  URL = {http://www.cs.cmu.edu/~yiming/papers.yy/icml97.ps.gz},
  ANNOTE = {The paper presents five different feature selection methods: Document Frequency, Information Gain, Mutual Information, CHI, and Term Strength. These methods were tested with both kNN and LLSF.}
}

@INPROCEEDINGS{yang99reexamination,
  AUTHOR = {Yiming Yang and Xin Liu},
  TITLE = {A re-examination of text categorization methods},
  BOOKTITLE = {Proc. ACM SIGIR},
  PAGES = {42--49},
  ADDRESS = {Berkley},
  YEAR = {1999},
  URL = {http://www.cs.cmu.edu/~yiming/papers.yy/sigir99.ps.gz},
  ANNOTE = {A good and concise introduction to the theory and results of somewhat recent development in text categorization. Yang and Liu experimented SVM, kNN, LLSF, Neural Networks and Naive Bayes.}
}