%0 Journal Article %1 FT001 %A Moens, M. %A Dumortier, J. %D 2000 %J Information Processing and Management. %K Automatic categorization, indexing, learning. machine text %N 6 %P 841-861 %T Text categorization: the assignement of subject descriptors to magazine articles. %V 36

@article{FT001, added-at = {2009-06-22T10:05:09.000+0200}, author = {Moens, M. and Dumortier, J.}, biburl = {https://www.bibsonomy.org/bibtex/2b315b4567779da10435cef7b5e310c2c/n.nanas}, interhash = {4beb4d48343706921ef469706d66bb77}, intrahash = {b315b4567779da10435cef7b5e310c2c}, journal = {Information Processing and Management.}, keywords = {Automatic categorization, indexing, learning. machine text}, number = 6, pages = {841-861}, qnote = {Very good paper for what I am looking. It covers most of the issues I am trying to address. The feature selection algorithm they are using seems simple but not so effective. They are using term frequency divided with the max term frequency for each one of the documents. I will have to test it. They are also using the x2 algorithm with the best results and argue that performs well also with a limited set of positive examples which is not what I know from foundations of statistical NLP. Very rich in references as well. Another problem with x2 is that it is also using negative examples.}, timestamp = {2009-06-23T10:19:15.000+0200}, title = {Text categorization: the assignement of subject descriptors to magazine articles.}, volume = 36, year = 2000 }