Tuesday, June 23, 2009
Good news from ICTIR 2009
A paper by Fabrizio Sebastiani and me, submitted to ICTIR 2009 has been accepted for publication.
The paper title is “Training Data Cleaning for Text Classification”, and discusses three different techniques for performing training data cleaning in the context of supervised learning for text classification.
Training data cleaning consists in devising ranking functions that sort the original training examples in terms of how likely it is that the human annotator has misclassified them, thereby providing a convenient means for the human annotator to revise the training set so as to improve its quality.
Add comment
Fill out the form below to add your own comments