A Borda count for collective sentiment analysis

Abstract

Sentiment analysis assigns a positive, negative or neutral polarity to an item or entity, extracting and aggregating individual opinions from their textual expressions by means of natural language processing tools. In this paper we observe that current sentiment analysis techniques are satisfactory in case there is a single entity under consideration, but can lead to inaccurate or wrong results when dealing with a set of multiple items. We argue in favor of importing techniques from voting theory and preference aggregation to provide a more accurate definition of the collective sentiment over a set of multiple items. We propose a notion of Borda count which combines individuals’ sentiment with comparative preference information, we show that this class of rules satisfies a number of properties which have a natural interpretation in the sentiment analysis domain, and we evaluate its behavior when faced with highly incomplete domains.