Handling subjective information through augmented (fuzzy) computation

Abstract

Since it can result in significant benefits for a company or an individual, the inclusion of information extracted from subjective social media content into a decision making process is becoming a more frequent activity. However, such benefits are usually linked to the usability of the extracted information, which, among other aspects, depends on the reliability of its source. In this regard, people whose understandings of a topic are alike to the understanding possessed by an information seeker can be considered fairly reliable information sources. Hence, we propose a novel technique for detecting social media users with whom an information seeker shares a similar understanding of a given topic. Through this technique, posts on social media are digested to build a kind of database consisting of augmented Atanassov fuzzy sets, or AAIFSs for short, each resembling a collection of experience-based evaluations given by a particular source with respect to a given topic. Since such AAIFSs can be used in comparisons in which not only the extents but also the contexts of those evaluations are taken into account for computation, extracting more reliable (and usable) information is possible. An illustrative example shows how the proposed technique works and how it can help to detect sources having a common understanding of a topic.