- VIGÍA, Technologies for the Exploration of Purpose-Driven Communities VIGÍA, Technologies for the Exploration of Purpose-Driven Communities
- Committee Elogia Media S.L.
- Supporting entities Gradiant
- Funding PID
- Opening / Ending January 2017 / December 2018
The purpose of VIGÍA is to provide a set of technologies to assist in the exploration of information from large communities attending to the content generated by the users within.
To improve analysts’ experience, VIGÍA focuses on the visualization of large graphs of communities (over a scalable backend that facilitates the generation of such graphs, not supported by current technologies), providing different levels of abstraction based on the analysis of the contents generated by the users -obtained from both social networks and the web (news, blogs, RSS). VIGÍA proposes different content analysis techniques, based on deep learning technologies for Natural Language Processing (NLP):
- Entities detection related to a given context, adding extra information from external open data sources. In this context, an entity must be understood as every concept whose existence is perceptible.
- Events detection, grouping contents by a non supervised approach, using the semantics of the text to organize contents according to the event they refer to, not regarding the moment in which they occur.
- Sentiment analysis, inferring subjective perceptions of public opinion about events related to entities.
As a use case, the resulting technologies will be oriented to the analysis of communities for online marketing applications (such as identifying needs regarding a product, finding communities willing to accept innovative products, etc.). However, VIGÍA aims to be also a technological solution for other sectors such as the financial sector, where graph analysis is applied for the detection of anomalous user behavior, or cybersecurity, in order to identify suspicious transactions between users in the deep web.
The project requires a strong initial investigation. The proposed technologies are very close to the state of the art and addressed mainly in academic environments. The main work of GRADIANT is the research of the technologies to be applied, the study of the literature and the research/selection/development of algorithms.