The Feature Difference Coefficient: Classification by Means of Feature Distributions

The Feature Difference Coefficient: Classification by Means of Feature Distributions

Title: The Feature Difference Coefficient: Classification by Means of Feature Distributions
Authors:  Ulli Waltinger and  Alexander Mehler
Pub/Conf:  Text Mining Services (TMS), March 23-25, Leipzig, 2009

Abstract:
This paper presents a model of text classification using feature frequency distribution. The proposed algorithm offers not only sensitivity to linguistic but also to structure features and calculates a unified fingerprint for each category. Classifi cation is done by finding the closest match to prelearned models using a simple distance metric. The approach will be evaluated against three different classification scenarios. Language identification, text classification based on the Reuters corpus and web genre classification.

BibTeX:

@inproceedings{1904094,
  author       = {Waltinger, Ulli and Mehler, Alexander},
  language     = {English},
  series       = {Proceedings of Text Mining Services (TMS), March 23-25, 
                  Leipzig, Germany},
  title        = {The Feature Difference Coefficient: 
                 Classification by Means of Feature Distributions},
  year         = {2009},
}

PDFBibTeX