Big Data Quality Metrics for Sentiment Analysis Approaches

In a world increasingly connected, and in which information flows quickly and affects a very large number of people, sentiment analysis has seen a spectacular development over the past ten years. This is due to the fact that the explosion of social networks has allowed anyone with internet access to publicly express his opinion. Moreover, the emergence of big data has brought enormous opportunities and powerful storage and analytics tools to the field of sentiment analysis. However, big data introduces new variables and constraints that could radically affect the traditional models of sentiment analysis. Therefore, new concerns, such as big data quality, have to be addressed to get the most out of big data. To the best of our knowledge, no contributions have been published so far which address big data quality in SA throughout its different processes. In this paper, we first highlight the most important big data quality metrics to consider in any big data project. Then, we show how these metrics could be specifically considered in SA approaches and this for each phase in the big data value chain.