Robust processing of noisy web-collected data

Jelke Bloem, Michaela Regneri, Stefan Thater; Proceedings of KONVENS 2012 (Main track: poster presentations), pp. 189-193, September 2012.

Abstract

Subject of this paper is a fine-grained multi-level annotation model to enhance opinion detection in German blog comments. Up to now, only little research deals with the fine-grained analysis of evaluative expressions in German blog comments. Therefore, we suggest a multi-level annotation model where different linguistic means as well as linguistic peculiarities of users formulation and evaluation styles in blog comments are considered. The model is intended as a basic schema for the annotation of evaluative expressions in blog data. This annotation provides suitable features for implementing methods to automatically detect user opinions in blog comments.

[pdf] [bibtex]