Better electrocardiograms thanks to research by Dieter De Paepe

(28-06-2021) In his doctoral research ‘Insight Mining in Time Series Data with Applications for Anomaly Detection', Dieter De Paepe is looking for techniques to optimise the analysis of time series (e.g. electrocardiograms).

As a result of the many success stories of early adopters, many companies now recognise the value of data collection and analysis. Where the emphasis used to be on keeping as wide a range of data as possible, more and more companies are now choosing to collect detailed time series in the hope of gaining new insights.

Common examples of time series include stock market prices or medical signals such as electrocardiograms.

Time series capture a certain property over time and thus allow us to follow how a certain property evolves. The time dimension in time series provides a new angle from which to view data and gain insights. For example, it is quite easy for us, as humans, to visually notice that a time series shows certain patterns or deviations. However, this human approach fails when the data is too extensive or patterns are less pronounced.

"In my thesis, I propose several techniques that allow to visualise, analyse and detect anomalies in time series. Thanks to my research, time series can be further optimised", concludes Dieter De Paepe.

About Dieter De Paepe

In 2010, Dieter obtained the degree Master of Science in Computer Science Engineering at Ghent University, with a specialisation in software engineering. Once graduated, he started working as a Talent Planet consultant for Thomson Reuters, where he worked on trademark protection and research. The daily commute to Antwerp was a bit of a draw, so in 2011 he decided to start working at Comsof in Ghent, an old spin-off of Ghent University. Dieter joined Comsof at the right time to help develop their first network-planning software, and quickly grew into the role of software architect. In 2015 he was in need of new challenges and started working in the semantic web group of IDLab within UGent. Here, Dieter got acquainted with the different research fields within IDLab. After his first publications and advising the Flemish government with the OSLO initiative, Dieter switched to the Predict research group within IDLab, which conducts research into applied machine learning.

Dieter De Paepe

Contact: Dieter De Paepe


Editor: Jeroen Ongenae - Final editing: Ilse Vercruysse - Illustrator: Roger Van Hecke