LIAAD Artificial Intelligence and Decision Support Lab of INESC TEC João Mendes Moreira
Synopsis Decision support Business Intelligence Fundamental Research
Decision Support For Public Transport Planning For Robotic Soccer For Maize Production For Human Activity recognition For Online Advertising Management
For public transport planning -Buses Main idea: Using data from the actual services to improve the planning Research topics Predicting travel times Validating bus schedules Bus bunching detection Recommending actions for bus bunching prevention (ongoing) Main techniques/methods involved: Regression methods SVM, PPR, Random forest Ensemble learning Classification methods Sequence classification Distribution rules Time series clustering Consensual clustering
For public transport planning -Buses People involved: Alípio Jorge, PhD (LIAAD); Carlos Ferreira, PhD student (LIAAD); Carlos Soares, PhD (UESP); Eva Duarte, MSc (UMinho); Luís Matias, PhD student (LIAAD); João Gama, PhD (LIAAD); Jorge Freire de Sousa, PhD (UGEI); Orlando Belo, PhD (UMinho); Paulo Azevedo, PhD (UMinho). Publications: 1 international journal paper: Intelligent Data Analysis; 15 conference papers; 1 national journal paper.
For public transport planning -Taxis Main idea: Using data from the actual services, to recommend the best taxi stand to go to after a passenger drop-off Research topics Predicting taxi-passenger demand Recommending the taxi-stand Main techniques/methods involved: Time-series forecasting methods on data streams Ensemble learning on data streams
For public transport planning -Taxis People involved: João Gama, PhD (LIAAD); Luís Damas, PhD (GeoLink); Luís Matias, PhD student (LIAAD); Michel Ferreira, PhD (FCUP); Ricardo Fernandes (GeoLink). Publications: 1 international journal paper: IEEE Transactions on Intelligent Transportation Systems; 5 conference papers.
For robotic soccer Main idea: To adapt the way our team plays according to how our opponent is playing Research topics Select the set of statistics to better characterize how the opponent team is playing To recommend strategies according to how the opponent team is playing Main techniques/methods involved: Feature selection Interpretable regression methods Recommending systems (model-based vs memory-based)
For robotic soccer People involved: Daniel Castro Silva, PhD (LIACC); Daniel Castelão, MSc; Israel Costa, PhD (FADEUP); João Portela, MSc; Júlio Garganta, PhD (FADEUP); Luís Paulo Reis, PhD (LIACC); Pedro Henriques Abreu, PhD (UCoimbra). Publications: 3 international journal papers: Intelligent Data Analysis International Journal of Computational Intelligence Systems European Journal of Sport Science
For maize production Main idea: To relate ear and field characteristics with maize production Research topics To identify the ear and field characteristics with influence in maize production To identify a criteria for best ear contests according to its potential productivity Main techniques/methods involved: Interpretable regression methods Instance ranking Evaluation measures for instance ranking
For maize production People involved: Arnel R. Hallauer, PhD (Iowa State University) E. Andrade J.P.N. Santos J.P.P. Santos M. Mota Maria Carlota Vaz Patto, PhD (ITQB-UNL) Pedro Mendes-Moreira, PhD student (ESAC) Silas E. Pêgo, PhD (INIAV) Publications: 1 international journal paper: Maydica 1 conference paper
For human activity recognition Main idea: To recognize human activities using sensors embedded in mobile phones Main techniques/methods involved: Feature construction Online classification Online semi-supervised classification People involved: Alexandre Lopes, MSc student João Gama, PhD (LIAAD) Nuno Silva, PhD student Paulo Menezes, PhD (ISR UCoimbra) Ricardo Cachucho, MSc student Publications: 2 conference papers
For online advertising management Main Idea: Managing online advertising campaigns for ShiftForward Research topics: Ranking users Predicting future sequences of values Mani techniques/methods involved: Instance ranking Instance-based regression People involved: Hugo Sereno Ferreira, PhD João Azevedo, MSc Jorge Silva, MSc student Pedro Borges, MSc student Rui Gonçalves, MSc
Data Mining Applications (yet ) Educational data mining (with Carlos Soares, PhD UESP and Pedro Strecht Ribeiro, PhD student) Strategies to thwart students' attrition Servers network analysis (with João Ponte, MSc student and Ricardo Morla, PhD) Outlier detection (offline and online) FEUP-WeDo project (with André Fernandes, MSc; Carlos Soares, PhD UESP; João Gama, PhD LIAAD, Rui Sarmento, MSc and WeDo) Big Data from telecommunications Distributed clustering Map-reduce OLAP Outlier detection
Business Intelligence UP Developing Data Marts for the University of Porto (LIAAD, UESP & UGEI) CoReNeT Customer-ORiented and Eco-friendly NETworks for healthy fashionable goods (2010-2013) An European Project Development of a BI system for Bivolino, a Belgian company of customized shirts 1 conference paper PT21 Powered Textiles Século 21 Funded: ERDF through the COMPETE Program and by QREN Development of a BI system for an intelligent shop ADIRA Defining Key Performance Indicators for the commercial department with Luís Machado, MSc student ALERT BI from unstructured data With Carlos Soares, PhD - UESP, Hélder Quintela, MSc Alert, Ricardo Pereira, MSc student
Fundamental research Ensemble Learning: Ensemble methods for regression with Alípio Jorge, PhD LIAAD; Carlos Soares, PhD UESP; Jorge Freire de Sousa, PhD UGEI Publications: 1 international journal paper: ACM Computing Surveys 1 conference paper Metalearning for Dynamic Integration in Ensemble Methods with Carlos Soares, PhD UESP and Fábio Pinto, PhD student LIAAD Ensemble learning for binary classification with João Gama, PhD, Luís Matias, PhD student and Pavel Brazdil, PhD all from LIAAD Publications: 1 conference paper Instance Learning: Interpretable Instance Ranking Models (pairwise vs. listwise approaches) with João Brito, MSc student