Post date: 11-ene-2018 14:50:51
Title: Online data-based topology estimation: applications in industrial and urban environments
Date: 8-Nov-2017, 13:00
Aula: 303 (Aulario III)
Material: póster_pdf
Speaker: Luis Miguel López Ramos (Univ. Agder Norway)
Abstract:
An important problem in data sciences pertains to inferring causal interactions among a collection of time series: these can be modelled as a so-called causality graph. This talk will cover identification of causality graphs from data, by means of vector autoregressive (VAR) modelling. To exploit the sparse connectivity of causality graphs, the proposed estimators minimize a group-Lasso regularized functional.
To cope with real-time applications, big data setups, and possibly time- varying topologies, we propose online algorithms that recover the sparse coefficients when observations are received sequentially.
The inferred graphs are useful for several applications, including forecasting, system identification, data completion and compression of time series. These applications will be illustrated in two real test cases: one from industry (oil and gas) and an urban water distribution network.
Short bio: I received the PhD degree in Telecommunication Engineering from Rey Juan Carlos University in 2016. Since 2017 I have been working as a post- doctoral research fellow in the Wisenet group in the University of Agder (Grimstad, Norway). My research focuses on optimization techniques applied into statistical signal and data processing, wireless communications, and energy management.