Paper: Developing Robust Models for Favourability Analysis

ACL ID W11-1706
Title Developing Robust Models for Favourability Analysis
Venue Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis
Year 2011

Locating documents carrying positive or neg- ative favourability is an important application within media analysis. This paper presents some empirical results on the challenges fac- ing a machine-learning approach to this kind of opinion mining. Some of the challenges in- clude: the often considerable imbalance in the distribution of positive and negative samples; changes in the documents over time; and ef- fective training and quantification procedures for reporting results. This paper begins with three datasets generated by a media-analysis company, classifying documents in two ways: detecting the presence of favourability, and as- sessing negative vs. positive favourability. We then evaluate a machine-learning approach to automate the classification process. We ex- plore the effect of u...