What is it about?
Insights toward complex disease, such as diabetic complications, can come from detailed data analysis, identifying and grouping patients with similar clinical characteristics, and investigating their response to the therapeutic treatment. In this research, we focus on diabetic kidney disease, analyzing time evolution of kidney filtering efficiency, quantified via the eGFR. We first cluster eGFR time series (trajectories) according to a shape-similarity criterion. Then, , we sonify them: we map each value of a trajectory into a sound, and each trajectory becomes a melody and a collection of values in Hertz. We finally build clusters based on Hertz values, comparing them with shape-based ones.
Featured Image
Photo by Robina Weermeijer on Unsplash
Why is it important?
Sonification constitutes a valuable tool to enrich data understanding, providing an additional sensory dimension. Sonification is more and more used also in the biomedical domain. In our research, we propose a sonification in precision medicine. The insights gained from sounds can help create a more complete and faceted representation of phenomena.
Perspectives
Read the Original
This page is a summary of: Trajectory-based and sound-based medical data clustering, August 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3535508.3545102.
You can read the full text:
Contributors
The following have contributed to this page