Publikationen

zurück
F. Spechtenhauser,  R. Hronsky,  T. MöllerH. Piringer (2018)

Towards a Visual Analytics Pipeline for the Analysis of Recurring Patterns in Time Series Data

Wo ist die Publikation erschienen?

VIS 2018 (Poster)

Abstract

The analysis of recurring patterns in time series plays a key role in many application domains, for example healthcare monitoring and predictive maintenance. This poster describes ongoing work towards a visual analytics pipeline for a holistic analysis of recurring temporal patterns, including segmentation, alignment, clustering, feature extraction, comparison and downstream tasks such as monitoring. The focus of this poster is on a novel interactive tool that covers parts of this pipeline, i.e., cluster analysis, interactive feature specification and extraction, and comparison of patterns. The work is motivated by the analysis of data from industrial batch production and was designed in close cooperation with domain experts in industrial quality management. The proposed dashboard is a first step towards a holistic visual analytics pipeline which is generally applicable for the analysis of recurring patterns in time series data.

Forschungsthemen