Abstract
With the ever-increasing number of social media messages posted daily, millions of users express opinions on various subjects, including opinions concerning the characteristics of products and services that they have already bought or they intend to buy in the near future. Accurately knowing the opinions of such a large number of users in near real time would be invaluable for the companies marketing those products. Thus, in the present paper, we propose an approach based on Semantic Web technologies, natural language processing and machine learning for accurately analysing the social media messages posted on Twitter. Compared to existing approaches, which mainly focus on determining the opinion of the user concerning the entire product, the approach proposed in the present paper offers deeper insights, by taking into consideration the fact that a user might have different and sometimes even contradictory opinions concerning the various characteristics of a single product. We start by creating an ontology for representing the relationships between the products and their characteristics, ontology that is also used for performing named entity recognition, given the fact that various users can employ different terms for referring to the same concept. The ontology is afterwards used in order to filter from the huge number of tweets published every minute only the ones that can prove relevant for the analysis. In the next step, aspect-based sentiment analysis is employed in order to determine the sentiment expressed by the social media user regarding one or several characteristics of the analysed product. The results of the analysis are stored as semantically structured data, thus making it possible to fully exploit the possibilities offered by Semantic Web technologies, such as inference and accessing the vast knowledge in Linked Open Data, for further analysis.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have