openHSU (HSU Hamburg)
Not a member yet
7291 research outputs found
Sort by
Technologien und Architekturen
No full textDer Beitrag beschreibt aus technischer Perspektive das Vorgehen bei der Implementierung des Lehr-/Lernpakets ComDigiS*, einem Angebot zur Diagnose und zielgerichteten Vermittlung digitaler Kompetenzen von Studierenden. Der Beitrag beginnt mit einer Darstellung der Rahmenbedingungen, innerhalb derer der Entwicklungsprozess stattfand. Im weiteren Verlauf erfolgt eine Betrachtung der Erhebung, Kategorisierung und Gewichtung der Anforderungen verschiedener Interessengruppen. In Bezug auf die technischen Aspekte werden verschiedene Optionen, Standards und Entwicklungswerkzeuge vorgestellt und hinsichtlich projektbezogener Kriterien unterschieden. Außerdem wird ein Überblick der daraus abgeleiteten Systemarchitektur gegeben, wobei der Schwerpunkt auf den Interaktionen zwischen den Softwarekomponenten liegt. Die Projektergebnisse werden abschließend im Rahmen der Evaluation von ComDigiS* anhand verschiedener Methoden und Zielgruppen überprüft.
Die Ergebnisse des Projekts dienen als Grundlage für die Entwicklung weiterer Bildungsformate und werden quelloffen (Open Source) und als Open Educational Resources (OER) zur Verfügung gestellt.Vo
What Cosco’s Hamburg deal can tell us about Europe’s “rules-first” approach to China
No full textBeitrag im Blog European Politics and Policy (EUROPP)Vo
Evaluating imputation techniques for short-term gaps in heart rate data
No full textRecent advances in wearable technology have enabled the continuous monitoring of vital physiological signals, essential for predictive modeling and early detection of extreme physiological events. Among these physiological signals, heart rate (HR) plays a central role, as it is widely used in monitoring and managing cardiovascular conditions and detecting extreme physiological events such as hypoglycemia. However, data from wearable devices often suffer from missing values. To address this issue, recent studies have employed various imputation techniques. Traditionally, the effectiveness of these methods has been evaluated using predictive accuracy metrics such as RMSE, MAPE, and MAE, which assess numerical proximity to the original data. While informative, these metrics fail to capture the complex statistical structure inherent in physiological signals. This study bridges this gap by presenting a comprehensive evaluation of four statistical imputation methods, linear interpolation, K Nearest Neighbors (KNN), Piecewise Cubic Hermite Interpolating Polynomial (PCHIP), and B splines, for short term HR data gaps. We assess their performance using both predictive accuracy metrics and statistical distance measures, including the Cohen Distance Test (CDT) and Jensen Shannon Distance (JS Distance), applied to HR data from the D1NAMO dataset and the BIG IDEAs Lab Glycemic Variability and Wearable Device dataset. The analysis reveals limitations in existing imputation approaches and the absence of a robust framework for evaluating imputation quality in physiological signals. Finally, this study proposes a foundational framework to develop a composite evaluation metric to assess imputation performance.Vo
Die Rolle der Kommunikation zwischen Führungskräften und Mitarbeitern in der Gesundheitsorientierten Führung
No full textThis article in the journal "Gruppe. Interaktion. Organisation." investigates whether the quantity and quality of communication between employees and their leaders are related to their leaders’ StaffCare. StaffCare, a key component of the Health-oriented Leadership concept, reflects leaders’ commitment to promoting health and their awareness of employees’ needs. Previous studies have mainly focused on several job demands and resources that may influence leaders’ StaffCare, while the role of communication between leaders and followers has received less attention. This study examines communication factors on a dyadic level. Study 1 was designed as a two-wave study with two measurement points two months apart. The online survey was conducted across various industries and companies in Germany. Hierarchical regression analyses of N = 320 employees show that frequency, communication barriers, and general informal communication are significant predictors of StaffCare. Study 2 was designed as a cross-sectional online survey conducted within an international pharmaceutical company in Germany. It examines informal communication, particularly SmallTalk and DeepTalk, as well as factors such as relationship tenure and stigma toward mental health, defined as negative attitudes and reactions towards psychological strain or illness. For this analysis, only non-leadership employees were considered (N = 199). Results confirm that both SmallTalk and DeepTalk have a significant influence on leaders’ StaffCare. Relationship tenure moderates the relationship, with long-term employees benefiting more from high-quality interactions. Perceived stigma toward mental health is negatively associated with StaffCare; however, the expected interaction effect with DeepTalk was not significant. This study extends the field of Health-oriented Leadership by identifying new antecedents of StaffCare. The findings underscore the importance of reducing workplace stigma toward mental health to create a health-supportive environment and suggest that leaders encourage both casual and in-depth conversations with employees. Additionally, leadership communication strategies should consider employees’ relationship tenure, as long-term employees gain more from these high-quality interactions.Vo
Introduction of soft-switching loss determination to behavioral modeling techniques
No full textThis paper addresses the fundamentals of a behavioral model for a half bridge, designed to accurately represent switching losses at behavioral simulation level. In particular, it extends the classification of transistor modeling and examines the concept of mixed switching as an encompassing term for incomplete zero voltage switching. Finally, the paper analyzes the resulting loss mechanisms associated with various switching behaviors in a half-bridge topology.Vo
Konsultative Bürgerräte auf kommunaler Ebene
No full textLicensed under CC BY-SA-4.0 (https://creativecommons.org/licenses/by-sa/4.0/legalcode)Laut dem Bericht „Bürgerräte in Deutschland“ finden rund 80% aller Bürgerräte auf kommunaler Ebene statt. Dies wirft die Frage auf, wie Länder und Kommunen einen rechtssicheren Rahmen für ihren Einsatz schaffen können. Die kommunale Selbstverwaltungsgarantie bietet hier Chancen, auch wenn ihr Verhältnis zum Demokratieprinzip sowie Kompetenzüberlegungen zu beachten sind.Vo
An improved simulation methodology for nanoparticle injection through aerodynamic lens systems
No full textAll article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).Aerosol injectors applied in single-particle diffractive imaging experiments demonstrated their potential in efficiently delivering nanoparticles with high density. Continuous optimization of injector design is crucial for achieving high-density particle streams, minimizing background gas, enhancing x-ray interactions, and generating high-quality diffraction patterns. We present an updated simulation framework designed for the fast and effective exploration of the experimental parameter space to enhance the optimization process. The framework includes both the simulation of the carrier gas and the particle trajectories within injectors and their expansion into the experimental vacuum chamber. A hybrid molecular-continuum-simulation method [direct simulation Monte Carlo (DSMC)/computational fluid dynamics (CFD)] is utilized to accurately capture the multi-scale nature of the flow. The simulation setup, initial benchmark results of the coupled approach, and the validation of the entire methodology against experimental data are presented. The results of the enhanced methodology show a significant improvement in the prediction quality compared to previous approaches.Vo
Computational paralinguistic and phonetic approaches for perceived leadership detection
No full textThis dissertation investigates the nexus of speech features and perceived transformational leadership through computational paralinguistic and phonetic approaches across three studies, bridging leadership theory and vocal.
Study 1, an integrative review, leverages computational advancements to explore acoustic features like pitch, jitter, and formant dispersion beyond human perception. It synthesizes early research with modern tools like Mel-Frequency Cepstral Coefficients (MFCCs), showing how lower pitch and features like speech pauses predict dominance and charisma (Cullen & Harte, 2018). This sets the stage for empirical analyses using machine learning to dissect vocal cues.
Study 2 employs a computational paralinguistic approach, analyzing 122 speakers’ recordings by evaluated by 122 raters via the German Multifactor Leadership Questionnaire (MLQ)—with OpenSMILE. Focusing on fundamental frequency, intensity, and voicing probability, it uses Sequential Minimal Optimization (SMO) regression in WEKA. Results highlight fundamental frequency’s predictive power for inspirational motivation (R²=0.31) and idealized influence (R²=0.45), with intensity driving individualized consideration (R²=0.35). This approach quantifies paralinguistic features’ impact, revealing their nuanced roles across leadership dimensions.
Study 3 shifts to a phonetic approach, using Praat on the same dataset to extract fundamental frequency (F0), intensity, speech duration, and formants (F1-F5). SMO regression identifies speech duration as key for individualized consideration (R²=0.40) and F0 for inspirational motivation (R²=0.43). The implications amplify these findings: a wider F0 range and steeper slopes enhance dynamism and charisma; lower F1, F2 frequencies and narrower F3, F4 bandwidths boost authoritative resonance and clarity; longer duration and pauses, paired with slower rates, project control; and dynamic intensity modulation strengthens emotional impact. These phonetic insights complement Study 2’s paralinguistic focus, offering a dual-lens framework.
The dissertation integrates explainable AI (XAI) to balance predictive accuracy with interpretability, linking computational paralinguistic features (e.g., voicing probability) and phonetic traits (e.g., formant bandwidths) to psychological constructs like enthusiasm and authority. The paralinguistic approach excels in broad feature extraction, while the phonetic method provides granular physiological insights, together advancing psychoacoustics and leadership studies. This synergy enables practical vocal optimization- varying pitch, modulating intensity, and pacing delivery- for authentic leadership projection in business contexts, demonstrating the power of computational and phonetic methodologies in decoding vocal influence.Diese Dissertation erforscht die Verbindung zwischen Sprachmerkmalen und wahrgenommener transformationaler Führung durch computationelle paralinguistische und phonetische Ansätze in drei Studien, die Führungstheorie und stimmlichen Ausdruck verknüpfen.
Studie 1, ein integrativer Reviewansatz, nutzt rechentechnische Fortschritte, um akustische Merkmale wie Tonhöhe, Jitter und Formantdispersion jenseits menschlicher Wahrnehmung zu analysieren. Sie kombiniert frühe Forschung mit Tools wie Mel-Frequency Cepstral Coefficients (MFCCs), um zu zeigen, wie tiefere Tonlagen und Merkmale wie Sprechpausen Dominanz und Charisma vorhersagen (Cullen & Harte, 2018), und legt die Grundlage für empirische Untersuchungen mit maschinellem Lernen.
Studie 2 wendet einen computationelle paralinguistischen Ansatz an, indem sie Aufnahmen von 122 Sprechern – bewertet von 122 Personen via deutscher Version des Multifactor Leadership Questionnaire (MLQ) – mit OpenSMILE analysiert. Fokussiert auf Grundfrequenz (F0), Intensität und Stimmhaftigkeitswahrscheinlichkeit, verwendet sie Sequential Minimal Optimization (SMO) Regressionsalgorithmus in Waikato Environment for Knowledge Analysis (WEKA). Ergebnisse zeigen die Vorhersagekraft der Grundfrequenz für inspirierende Motivation (R²=0,31) und idealisierten Einfluss (attributiert) (R²=0,45), während Intensität individuelle Berücksichtigung prägt (R²=0,35). Dieser Ansatz quantifiziert den Einfluss paralinguistischer Merkmale und deckt ihre nuancierten Rollen auf.
Studie 3 verfolgt einen phonetischen Ansatz mit Praat (Version 6.3.04) auf demselben Datensatz, extrahiert Grundfrequenz (F0), Intensität, Sprechdauer und Formanten (F1-F5). SMO-Regression identifiziert Sprechdauer als Schlüsselfaktor für individuelle Berücksichtigung (R²=0,40) und F0 für inspirierende Motivation (R²=0,43). Es vertieft diese Erkenntnisse: Ein größerer F0-Bereich und steilere Änderungen fördern Dynamik und Charisma; niedrigere F1- und F2-Werte steigern autoritative Resonanz und Klarheit; schmalere Formanten-Bandbreite bei F3 und F4 verbessern Resonanzklarheit; längere Gesamtlänge der Sprechdauer und Pausen mit langsamerer Sprechgeschwindigkeit projizieren Kontrolle; dynamische Intensitätskurve verstärkt emotionale Wirkung. Diese phonetischen Einsichten ergänzen die paralinguistischen Befunde aus Studie 2 und bieten einen dualen Analyseansatz.
Durch Integration von Prinzipien der erklärbaren künstlichen Intelligenz (XAI) vereint die Arbeit prädiktive Genauigkeit mit Interpretierbarkeit, indem sie paralinguistische Merkmale (z. B. Stimmhaftigkeitswahrscheinlichkeit) und phonetische Eigenschaften (z. B. Formanten-Bandbreite) mit psychologischen Konstrukten wie Enthusiasmus und Autorität verknüpft. Der paralinguistische Ansatz glänzt bei breiter Merkmalsextraktion, der phonetische bietet detaillierte physiologische Einblicke. Zusammen bereichern sie Psychoakustik und Führungsforschung und ermöglichen praktische Stimmoptimierung – durch Tonhöhenvariation, Intensitätsmodulation und gezielte Pausensetzung – für authentische Führungspräsenz im Geschäftskontext, was die Stärke dieser Methoden unterstreicht.Vo
Deep learning-based hypoglycemia classification across multiple prediction horizons
No full textThis work is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).Type 1 diabetes (T1D) management can be significantly enhanced through the use of predictive machine learning (ML) algorithms, which can mitigate the risk of adverse events like hypoglycemia. Hypoglycemia, characterized by blood glucose levels below 70 mg/dL, is a life-threatening condition typically caused by excessive insulin administration, missed meals, or physical activity. Its asymptomatic nature impedes timely intervention, making ML models crucial for early detection. This study integrates short- (up to 2h) and long-term (up to 24h) prediction horizons (PHs) within a single classification model to enhance decision support. The predicted times are 5-15 min, 15-30 min, 30 min-1h, 1-2h, 2-4h, 4-8h, 8-12h, and 12-24h before hypoglycemia. In addition, a simplified model classifying up to 4h before hypoglycemia is compared. We trained ResNet and LSTM models on glucose levels, insulin doses, and acceleration data. The results demonstrate the superiority of the LSTM models when classifying nine classes. In particular, subject-specific models yielded better performance but achieved high recall only for classes 0, 1, and 2 with 98%, 72%, and 50%, respectively. A population-based six-class model improved the results with at least 60% of events detected. In contrast, longer PHs remain challenging with the current approach and may be considered with different models.N
Gulf of America: rewriting the rules of public discourse
No full textDer Beitrag ist im Blog "The Loop : ECPR's Political Science Blog" erschienen.Scientific investigation into how disinformation affects democracy has never been more important. But autocrats and populists discredit such research, along with any journalism that challenges their worldview. Christoph Deppe describes how Trump’s second administration is changing the rules of communication – and manipulating public discourse.Vo