Full Consideration of Big Data Characteristics in Sentiment Analysis Context

Abstract

We live in a digital world where every interaction with Internet generates digital traces that reveal a lot of details about our thinking. The volume of these daily-generated traces increases exponentially creating massive loads of information, which represents a big part of a new world called Big Data. This big data is used by different type of projects to extract valuable information either to take marketing decisions, track specific behaviors or detect threat attacks. Sentiment analysis (SA) is one of the most active research areas relying on big data, even if their involvement would be of great added value. However, most of its applications, like other big data projects, consider only the volume, sometimes also the veracity, and completely ignore the rest of big data characteristics (Variety, Velocity, Value, Variability, and Visualization). In this paper, we focus on projecting big data characteristics in a sentiment context to get out the most of big data. Our main contribution consists of presenting and analyzing the most known sentiment analysis approaches and contributions that rely on big data by showing their main characteristics. Furthermore, we explain how SA applications should consider big data characteristics to be fully aligned with big data contexts.