Ralf Krestel

• About Me
• Publications
  • Book Chapters
  • Journal Articles
  • Conference Papers
    • ESWC 21
    • CHIIR 21
    • WI 20
    • CIKM 20
    • KI 20
    • ICWSM 20
    • JCDL 20a
    • JCDL 20b
    • TPDL 19
    • ICADL 18
    • ICDIM 18
    • JCDL 18a
    • JCDL 18b
    • NAACL 18
    • ECIR 18
    • ICDM 17
    • TPDL 17
    • NLDB 16
    • WI 15
    • KI 15
    • RECSYS 13
    • TPDL 13
    • HT 13
    • WI 11
    • KI 10
    • ECDL 10
    • WI 10
    • NLDB 10
    • LAWEB 09
    • RECSYS 09
    • WI 08
    • ASWC 08
    • LAWEB 08
    • LREC 08
    • RANLP 07
    • CanadianAI 07
  • Workshop Papers
  • Posters & Demos
  • Proceedings
  • Others
• Travels

HT 13

Generating Contextualized Sentiment Lexica based on Latent Topics and User Ratings

Abstract

Sentiment lexica are useful for analyzing opinions in Web collections, for domain-dependent sentiment classification, and as sub-components of recommender systems. In this paper, we present a strategy for automatically generating topic-dependent lexica from large corpora of review articles by exploiting accompanying user ratings. Our approach combines text segmentation, discriminative feature analysis techniques, and latent topic extraction to infer the polarity of n-grams in a topical context. Our experiments on rating prediction demonstrate a substantial performance improvement in comparison with existing state-of-the-art sentiment lexica.

Full Paper

HT13.pdf

Conference Homepage

HT 2013

BibTex Entry

@InProceedings{krestel-ht13, author = {Ralf Krestel and Stefan Siersdorfer}, title = {{Generating Contextualized Sentiment Lexica based on Latent Topics and User Ratings}}, booktitle = {{Conference on Hypertext and Social Media}}, series = {HT}, location = {Paris, France}, pages = {129--138}, month = {May 1--3}, year = {2013}, publisher = {ACM} }