Too Good To Be True: accuracy overestimation in (re)current practices for Human Activity Recognition

Andres Tello*, Victoria Degeler, Alexander Lazovik

*Bijbehorende auteur voor dit werk

OnderzoeksoutputAcademicpeer review

Samenvatting

Today, there are standard and well established procedures within the Human Activity Recognition (HAR) pipeline. However, some of these conventional approaches lead to accuracy overestimation. In particular, sliding windows for data segmentation followed by standard random k-fold cross validation, produce biased results. An analysis of previous literature and present-day studies, surprisingly, shows that these are common approaches in state-of-the-art studies on HAR. It is important to raise awareness in the scientific community about this problem, whose negative effects are being overlooked. Otherwise, publications of biased results lead to papers that report lower accuracies, with correct unbiased methods, harder to publish. Several experiments with different types of datasets and different types of classification models allow us to exhibit the problem and show it persists independently of the method or dataset.

Originele taal-2English
Titel2024 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2024
UitgeverijIEEE
Pagina's511-517
Aantal pagina's7
ISBN van elektronische versie979-8-3503-0436-7
ISBN van geprinte versie979-8-3503-0437-4
DOI's
StatusPublished - 23-apr.-2024
Evenement2024 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2024 - Biarritz, France
Duur: 11-mrt.-202415-mrt.-2024

Publicatie series

Naam2024 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2024

Conference

Conference2024 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2024
Land/RegioFrance
StadBiarritz
Periode11/03/202415/03/2024

Vingerafdruk

Duik in de onderzoeksthema's van 'Too Good To Be True: accuracy overestimation in (re)current practices for Human Activity Recognition'. Samen vormen ze een unieke vingerafdruk.

Citeer dit