Date created: | Last Updated:

Category: Project

Description: Nowadays, it is undoubtedly the growth and increase of social media usage like twitter, facebook or instagram as well as opinion portals about services like tripadvisor, booking or Holidaycheck where millions of users participate expressing their own opinions or sentiments about a certain topic. The companies are every day more conscious of the potential of analysing social media data to obtain valuable information that can beneficiate their marketing campaign, the acceptance of new products, etc. in order to increase their profits. Additionally, the analysis of social media is not privative of companies, and other social fields like journalists, politics, educational institutions, etc. can beneficiate from the opinion analysis about certain topics like for instance, the opinion about a certain law that will be approved in the govern, the sentiments around the remodelling or reform proposed, how polarized are the population about some political topics, etc. Examples of this type of analysis can be found in [Terán and Mancera, 2019] or [Diaz-Garcia et al., 2020] where the sentiment analysis was employed to inform about an electoral campaign, or for crime [Sharma, 2018], [Bolla, 2014] or radicalism detection [Fernandez and Alani, 2018]. Such is the interest in this field, that the European Union has provided an Action Plan in 2018 to “protect the Union’s democratic systems and combat disinformation”. For that, the EU has offered funding in different calls (e.g. under the topics: Evolving media landscapes and Europeanization, Behavioural, social and economic impacts of the outbreak response, and Culture, creativity and inclusive society) to fight against fake news and prevent disinformation in the society. However, traditional analysis techniques often fail when manipulating and processing such an enormous quantity of data that it is often unstructured and comes from heterogeneous sources. Therefore, SAPIENS pursues to develop new techniques and procedures capable of processing continuous data flows (i.e. streaming data) using distributive techniques under the umbrella of Big Data for the analysis of social media data. To achieve this objective, SAPIENS will investigate Natural Language Processing (NLP) tools jointly with Machine Learning (ML) and Data Mining (DM) techniques, incorporating Deep Learning (DL) algorithms to adequately process textual data from social media. Moreover, one of the weaknesses of the current proposals is the lack of a proper visualization that facilitates the inspection and interpretation of results. For this reason, SAPIENS will also incorporate advances in visualization techniques to facilitate the examination and results understanding. Additionally, all the procedures developed in SAPIENS will be applied to the analysis of social media from different perspectives: Sentiment analysis, Opinion mining and Misinformation; and will be validated by comparing them with the state of the art proposals.