Institut

Studium

Forschung


zur Startseite

Christian Scheible

Christian Scheible

???detail.Employee.Picture???Christian Scheible
Telefon +49 711 685-81386
Raum0.013
E-Mail
Anschrift
Universität Stuttgart
Institut für Maschinelle Sprachverarbeitung
Pfaffenwaldring 5b
70569 Stuttgart
Deutschland

I'm a 3rd year PhD student working mostly on Sentiment Analysis. My research interests include

  • Natural Language Processing with machine learning, in particular
    • generative models
    • bayesian models
    • sampling methods
    • graph-based methods
  • Sentiment Analysis using language modeling techniques

Lehre
Publikationen
  • Christian Scheible and Hinrich Schütze
    Unsupervised Sentiment Analysis with a Simple and Fast Bayesian Model using Part-of-Speech Feature Selection.
    Proceedings of the 1st Workshop on Practice and Theory of Opinion Mining and Sentiment Analysis, 2012.
    [pdf] [bib]
  • Christian Scheible and Hinrich Schütze
    Bootstrapping Sentiment Labels For Unannotated Documents With Polarity PageRank.
    Proceedings of the Eigth conference on International Language Resources and Evaluation (LREC), 2012.
    [pdf] [bib]
  • Florian Laws, Christian Scheible and Hinrich Schütze
    Active Learning with Amazon Mechanical Turk.
    Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2011.
    [pdf] [bib]
  • Sabine Schulte im Walde, Helmut Schmid, Wiebke Wagner, Christian Hying and Christian Scheible
    A Clustering Approach to Automatic Verb Classification incorporating Selectional Preferences: Model, Implementation, and User Manual.
    SinSpeC: Working Papers of the SFB 732 "Incremental Specification in Context", Volume 7, December 2010.
    [view]
  • Christian Scheible, Florian Laws, Lukas Michelbacher and Hinrich Schütze
    Sentiment Translation through Multi-Edge Graphs.
    Proceedings of the 23rd International Conference on Computational Linguistics (Coling), 1104-1112, 2010.
    [pdf] [bib]
  • Florian Laws, Lukas Michelbacher, Beate Dorow, Christian Scheible, Ulrich Heid and Hinrich Schütze
    A Linguistically Grounded Graph Model for Bilingual Lexicon Extraction.
    Proceedings of the 23rd International Conference on Computational Linguistics (Coling), 614-622, 2010.
    [pdf] [bib]
  • Christian Scheible
    Sentiment Translation through Lexicon Induction.
    Proceedings of the ACL Student Research Workshop (ACL SRW), 25-30, 2010.
    [pdf] [bib]
  • Christian Scheible
    An Evaluation of Predicate Argument Clustering using Pseudo-Disambiguation.
    Proceedings of the Seventh conference on International Language Resources and Evaluation (LREC), 2010.
    [pdf] [bib]
  • Beate Dorow, Florian Laws, Lukas Michelbacher, Christian Scheible and Jason Utt
    A Graph-Theoretic Algorithm for Automatic Extension of Translation Lexicons.
    Proceedings of the EACL Workshop on Geometrical Methods for Natural Language Semantics (GEMS), 2009.
    [pdf] [bib]
  • Sabine Schulte im Walde, Christian Hying, Christian Scheible and Helmut Schmid
    Combining EM Training and the MDL Principle for an Automatic Verb Classification incorporating Selectional Preferences.
    Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics (ACL), 2008.
    [pdf] [bib]

Undergraduate/Graduate Theses

  • Christian Scheible:
    Graph-Based Sentiment Translation,
    Diplomarbeit, 2010.
  • Christian Scheible:
    Evaluating a Verb Clustering Model Using Pseudo-Disambiguation,
    Studienarbeit, 2009.
Ressourcen

Software

Unsupervised Bayesian Naive Bayes

Python implementation

A Python implementation of unsupervised Bayesian Naive Bayes with Gibbs sampling as used in

Christian Scheible and Hinrich Schütze: Unsupervised Sentiment Analysis with a Simple and Fast Bayesian Model using Part-of-Speech Feature Selection. Proceedings of the 1st Workshop on Practice and Theory of Opinion Mining and Sentiment Analysis, 2012.

Available on github (zip) (tgz).

Hierarchical Bayes Compiler models

The experiments in the paper were conducted using models compiled with HBC: