The European programme IACOBUS granted the “Scientific Publications (Papers)” award to a team of researchers from INESC TEC’s Artificial Intelligence and Decision Support Laboratory (LIAAD), led by João Gama.
The awarded papers – “From Mobility Data to Habits and Common Pathways” and “A scalable saliency-based feature selection method with instance-level information” – were produced by INESC TEC researchers Thiago Andrade and Brais Cancela, together with Verónica Bolón Canedo, in partnership with the University of A Coruña, in Spain.
The first work presents an innovative approach of clustering algorithms to analyse individual mobility data (most frequent or important locations, for example), and to understand the pattern of individuals’ movements, the best routes, the time-frequency, and schedules, among others. The study can be applied to several areas, from destination forecasting to recommendation systems and vehicle sharing, based on the individuals’ habits.
The second paper focuses on developing a new saliency-based Feature Selection approach, which allows ranking the relevance of each instance-level attribute rather than concerning a set of examples. The algorithm explored in this work can be applied to any classification and regression problem, having been tested with different complex datasets. Moreover, the new approach can be used in different areas, like data analysis, semantic recognition, and classification of satellite images.
Both works reinforce the role that INESC TEC plays as an R&D institution in data mining and machine learning, presenting instrumental contributions to the state-of-the-art in these scientific fields.
By awarding this prize, the IACOBUS programme aims to promote cross-border scientific collaboration between the northern region of Portugal and the autonomous region of Galicia, in Spain. This programme also focuses on promoting scientific communication and dissemination activities through financial support to publish articles in scientific journals indexed in the Clarivate Analytics Journal Citation Report (JCR).
The INESC TEC researchers mentioned in this news piece are associated with IPP-ISEP, UP-FEUP and INESC TEC.