Data-Driven Discovery of Feature Groups in Clinical Time Series
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914151076462592 |
|---|---|
| author | Sergeev, Fedor Burger, Manuel Leshetkina, Polina Fortuin, Vincent Rätsch, Gunnar Kuznetsova, Rita |
| author_facet | Sergeev, Fedor Burger, Manuel Leshetkina, Polina Fortuin, Vincent Rätsch, Gunnar Kuznetsova, Rita |
| contents | Clinical time series data are critical for patient monitoring and predictive modeling. These time series are typically multivariate and often comprise hundreds of heterogeneous features from different data sources. The grouping of features based on similarity and relevance to the prediction task has been shown to enhance the performance of deep learning architectures. However, defining these groups a priori using only semantic knowledge is challenging, even for domain experts. To address this, we propose a novel method that learns feature groups by clustering weights of feature-wise embedding layers. This approach seamlessly integrates into standard supervised training and discovers the groups that directly improve downstream performance on clinically relevant tasks. We demonstrate that our method outperforms static clustering approaches on synthetic data and achieves performance comparable to expert-defined groups on real-world medical data. Moreover, the learned feature groups are clinically interpretable, enabling data-driven discovery of task-relevant relationships between variables. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_08260 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Data-Driven Discovery of Feature Groups in Clinical Time Series Sergeev, Fedor Burger, Manuel Leshetkina, Polina Fortuin, Vincent Rätsch, Gunnar Kuznetsova, Rita Machine Learning 68T07 (Primary), 62H30, 62M10, 62P10 (Secondary) I.2.6; I.5.3; J.3; G.3 Clinical time series data are critical for patient monitoring and predictive modeling. These time series are typically multivariate and often comprise hundreds of heterogeneous features from different data sources. The grouping of features based on similarity and relevance to the prediction task has been shown to enhance the performance of deep learning architectures. However, defining these groups a priori using only semantic knowledge is challenging, even for domain experts. To address this, we propose a novel method that learns feature groups by clustering weights of feature-wise embedding layers. This approach seamlessly integrates into standard supervised training and discovers the groups that directly improve downstream performance on clinically relevant tasks. We demonstrate that our method outperforms static clustering approaches on synthetic data and achieves performance comparable to expert-defined groups on real-world medical data. Moreover, the learned feature groups are clinically interpretable, enabling data-driven discovery of task-relevant relationships between variables. |
| title | Data-Driven Discovery of Feature Groups in Clinical Time Series |
| topic | Machine Learning 68T07 (Primary), 62H30, 62M10, 62P10 (Secondary) I.2.6; I.5.3; J.3; G.3 |
| url | https://arxiv.org/abs/2511.08260 |