Data and knowledge
Data science involves many actions on data. Data mining, also known as knowledge extraction from data, requires many steps, before and after learning. A major specificity of our team is not only to extract konwledge but also to use it, integrate knowledge in the learning process. Our research activities concern three areas.
We work with and learn from relational data, and more generally any data involving several "kinds" of objects. We study propositionalisation, ie. the automatic process to transform relational data into attribute-value data that can be used by any standard learner.
We are specialised in data representation and model data using relational models or graphs. We particularly consider their adaptations to temporal and spatial data.
Knowledge modelling and semantic technologies
A first topic concerns qualitative spatial reasoning, and case-based reasoning.
A second topic is on fuzzy reasoning using ontologies and on analogies between documents.
Database construction and data quality
Our activities include crowdsourcing, and the characterization and generation of suitable benchmarks.