Bahrololloomi, FarnodLuderschmidt, JohannesStaab, SergioKlein, MaikeKrupka, DanielWinter, CorneliaGergeleit, MartinMartin, Ludger2024-10-212024-10-212024978-3-88579-746-3https://dl.gi.de/handle/20.500.12116/45191In this work, we present a conceptual framework to determine the relevant recording time for nursing records, considering the effective detection and minimal interference with privacy. Our goal is to reduce the documentation burden, while ensuring compliance with the requirements of the General Data Protection Regulation (GDPR). We focus on data minimization and use a combination of speaker, context, and pronoun classification to accurately distinguish between nursing staff, patients, and visitors. Our work might indicate that when context and pronoun classification are used to identify patients, age classification becomes redundant. Furthermore, we address the challenges posed by non-native speakers in nursing homes, as language proficiency significantly affects the performance of language processing models. This work forms the basis for the automation of documentation processes in nursing homes.enSpeech-Based Activity RecognitionElectronic Health RecordGeneral Data Protection RegulationSpeaker IdentificationContext AnalysisNatural Language ProcessingText/Conference Paper10.18420/inf2024_321617-5468