Difference between revisions of "Melia et al AISTATS 2001"

From Cohen Courses
Jump to navigationJump to search
(Created page with '== Citation == Marina Melia and Jianbo Shi. 2001. A Random Walks View of Spectral Segmentation. In AISTATS 2001. == Online version == [http://www.stat.washington.edu/mmp/Paper…')
 
Line 9: Line 9:
 
== Summary ==
 
== Summary ==
  
This is an early and influential [[Category::paper]] that introduced the use of supervised learning for [[AddressesProblem::review classification]].  The authors used a corpus of [[UsesDataset::Pang Movie Reviews|1400 movie reviews]] that had been rated (by the authors) as positive or negative, and compared several approaches to predicting polarity.
+
This paper gives a general theoretical interpretation of a wide variety of spectral methods.  The authors first present the general framework of normalized cutsBegin with an index set <math>I</math>
 
 
# A hand-coded lexicon of polar wordsThis gave accuracy of around 70% (the dataset is balanced so random chance would be 50%).
 
# Off-the-shelf classifier learners ([[UsesMethod::Naive Bayes classifier learning|Naive Bayes]] and [[UsesMethod::Support vector machine classifier learning|SVM-lite]]) which performed well on topical text classification.  This gave accuracy in the high 70's and low 80's.
 
 
 
Some feature-engineering techniques suggested by the results in [[RelatedPaper::Turney, ACL 2002]] were explored, such as using phrases (bigrams) instead of unigrams, and using part of speech information, without major improvements in accuracy.
 
  
 
== Related papers ==
 
== Related papers ==

Revision as of 12:51, 4 February 2011

Citation

Marina Melia and Jianbo Shi. 2001. A Random Walks View of Spectral Segmentation. In AISTATS 2001.

Online version

Available on Marina Melia's Website

Summary

This paper gives a general theoretical interpretation of a wide variety of spectral methods. The authors first present the general framework of normalized cuts. Begin with an index set

Related papers