School of
Information Technology and Electrical Engineering

Speaker: Dr Aida Brankovic
Seminar Date: Thu, 10/10/2019 - 09:00
Venue: 47A-470; Sir James Foots building
Host: Prof Amin Abbosh

Seminar Type:  ITEE Research Seminar

Abstract: 

Today huge structured and unstructured data, consisting of thousands of variables or features, are available and used to extract the knowledge and provide an insight into future behaviour. It is therefore crucial to select the most relevant ones and combine them in order to obtain robust, reliable, and easily interpretable models. In that light Feature Selection (FS) aims precisely at selecting features that allow good discrimination among unseen samples of different classes or predicting the time sequence while avoiding the overfitting. In this talk, we will introduce a theoretical base for novel statistics-based and optimization techniques capable effectively to represent the knowledge and achieve robust prediction. To address problems with large scale feature sets, distributed optimisation scheme is proposed which additionally resulted in significantly reduced computational time. Experimental results on various benchmark datasets as well as real-life data are presented. This seminar will present this research being applied in the application fields of health, big data, energy and transportation.

Biography: 

Aida Brankovic received her BSc and MSc degree, in 2011 and 2013, respectively both in Automation, Control and Electronics from The University of Sarajevo. From February to September 2013 she worked as a teaching assistant at the Electrical Faculty of the University of Sarajevo, and from September to November 2013 as a research assistant in the MOVE research group of the Politecnico di Milano. In February 2018 she pursued her Ph.D. from Politecnico di Milano, Italy. Currently, she is working as a postdoctoral researcher at the University of Queensland. Her research interests include nonlinear model identification, statistical modelling, randomized algorithms and supervised machine learning and their real-world applications. Her research outcome includes 4 patents in energy and health areas, 1 simulation software package, 1 research grant and 10 journal and conference papers in published in the world-leading journals.