| 1. |
|
|
| 2. |
|
|
| 3. |
- Ouhaichi, Hamza, 1993-
(author)
-
A framework for designing and analyzing multimodal learning analytics systems
- 2024
-
Doctoral thesis (other academic/artistic)abstract
- The integration of technology in education offers transformative potential, especially with the advent of data-driven approaches that can personalize learning, support educators, and provide valuable insights into the learning process. Multimodal learning analytics (MMLA) holds remarkable promise within this context. By capturing and analyzing data from multiple sources—including video, audio, and digital interactions—MMLA systems offer a holistic view of learning experiences and the ability to tailor interventions in real time. This application has profound implications for understanding and enhancing learning experiences. However, the design of such sophisticated systems poses a significant challenge. Without conventional and field-tested frameworks, MMLA system development often remains self-driven and tailored to specific contexts, limiting both these systems’ broader adoption and full utilization. This thesis proposes a structured framework for designing MMLA systems across diverse educational contexts to address this fundamental challenge. The development of the framework followed a multifaceted methodology. In addition, action design research involving empirical studies, literature reviews, and expert interviews was employed to establish a set of foundational design considerations. The framework was then applied and refined within real-world educational settings. These included applications in the context of a globally distributed classroom and language acquisition environments. This practical application led to refinements that enhanced the framework’s adaptability and user-centric design. This thesis makes three key contributions: (1) a set of design considerations for MMLA systems, (2) a framework offering a structured guide for the design of MMLA systems, and (3) a conceptual system demonstrating the framework’s principles. The implications of this work are significant for researchers and stakeholders in MMLA, providing a foundation for future MMLA system development and ensuring more systematic and conventional design practices. This structured approach paves the way for broader adoption and integration of MMLA, ultimately enhancing educational outcomes and fostering personalized learning environments.
|
|
| 4. |
- Torén, Kjell, et al.
(author)
-
Restrictive Spirometric Pattern and Preserved Ratio Impaired Spirometry in a Population 50-64 Years.
- 2024
-
In: Annals of the American Thoracic Society. - 2329-6933 .- 2325-6621.
-
Journal article (peer-reviewed)abstract
- RATIONALE: Knowledge regarding prevalence and shared and unique characteristics of Restrictive spirometric pattern (RSP) and Preserved ratio impaired spirometry (PRISm) is lacking for a general population investigated with post-bronchodilator spirometry and computed tomography of the lungs.OBJECTIVES: To investigate shared and unique features for RSP and PRISm.METHODS: In the Swedish CArdioPulmonary bioImage Study (SCAPIS), a general population sample of 28,555 people aged 50 - 64 years (including 14,558 never-smokers) was assessed. The participants answered a questionnaire and underwent computed tomography of the lungs, post-bronchodilator spirometry, and coronary artery calcification score (CACS). Odds ratios (OR) with 95% confidence intervals (CI) were calculated using adjusted logistic regression. RSP was defined as FEV1/FVC≥0.70 and FVC<80%. PRISm was defined as FEV1/FVC≥0.70 and FEV1<80%. A local reference equation was applied.MEASUREMENTS AND RESULTS: The prevalence of RSP and PRISm were 5.1% (95% CI 4.9 - 5.4) and 5.1% (95% CI 4.8 - 5.3), respectively, with similar values seen in never-smokers. For RSP and PRISm, shared features were current smoking, dyspnea, chronic bronchitis, rheumatic disease, diabetes, ischemic heart disease (IHD), bronchial wall thickening, interstitial lung abnormalities (ILA), and bronchiectasis. Emphysema was uniquely linked to PRISm (OR 1.69, 1.36-2.10) vs 1.10 (0.84-1.43) for RSP. CACS≥300 was related to PRISm, but not among among never-smokers.CONCLUSIONS: PRISm and RSP have respiratory, cardiovascular, and metabolic conditions as shared features. Emphysema is only associated with PRISm. Coronary atherosclerosis may be associated with PRISm. Our results indicate that RSP and PRISm may share more features than not. This article is open access and distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).
|
|
| 5. |
- Veisi, Mohammad, et al.
(author)
-
jrfapp: A Python Package for Joint Inversion of Apparent S-Wave Velocity and Receiver Function Time Series
- 2024
-
In: Pure and Applied Geophysics. - 0033-4553 .- 1420-9136. ; 181:1, s. 65-86
-
Journal article (peer-reviewed)abstract
- Receiver function (RF) inversion is a well-established method to quantify a horizontally layered approximation of the S-wave velocity structure beneath a seismic station. It is well-known that the RF inverse problem is highly non-unique, and various tools such as joint inversion with other seismological observations exist that may overcome this problem. We present a joint inversion framework along with a Python package that implements the joint inversion of RF and the apparent S-wave velocity (VS,app). Our implementation includes a pseudo-initial model estimation, which helps address the inherent non-uniqueness of the joint inversion of RFs and VS,app. This implementation enhances the resolving power, enabling estimation of S-wave velocities with resolution approaching that of deep controlled source seismic methods. As an illustration, we showcase an example from a permanent station in the Makran subduction zone southeast of the Iranian Plateau and two other stations in the supplementary material. We compare our joint inversion results with several S-wave velocity models obtained through a deep seismic sounding profile and joint inversion of surface wave dispersion and RFs. This comparison shows that although we note a slightly lower sensitivity of our proposed method at greater depths (beyond 50 km), the method yields much better results for shallow structures. Our inversion code provides a powerful, accessible software package that has superior resolving power at shallow depth compared to RFs-surface wave inversion codes. Furthermore, the fact that only one data-derivative is used, makes this inversion code extremely easy to use, without the need for complementary datasets.
|
|
| 6. |
- Veisi, Mohammad, et al.
(author)
-
jrfapp: A Python Package for Joint Inversion of Apparent S-Wave Velocity and ReceiverFunction Time Series
- 2024
-
In: Pure and Applied Geophysics. - 0033-4553 .- 1420-9136.
-
Journal article (peer-reviewed)abstract
- Receiver function (RF) inversion is a well-estab-lished method to quantify a horizontally layered approximation ofthe S-wave velocity structure beneath a seismic station. It is well-known that the RF inverse problem is highly non-unique, andvarious tools such as joint inversion with other seismologicalobservations exist that may overcome this problem. We present ajoint inversion framework along with a Python package thatimplements the joint inversion of RF and the apparent S-wavevelocity (VS,app). Our implementation includes a pseudo-initialmodel estimation, which helps address the inherent non-uniquenessof the joint inversion of RFs and VS,app. This implementationenhances the resolving power, enabling estimation of S-wavevelocities with resolution approaching that of deep controlledsource seismic methods. As an illustration, we showcase anexample from a permanent station in the Makran subduction zonesoutheast of the Iranian Plateau and two other stations in the sup-plementary material. We compare our joint inversion results withseveral S-wave velocity models obtained through a deep seismicsounding profile and joint inversion of surface wave dispersion andRFs. This comparison shows that although we note a slightly lowersensitivity of our proposed method at greater depths (beyond50 km), the method yields much better results for shallow struc-tures. Our inversion code provides a powerful, accessible softwarepackage that has superior resolving power at shallow depth com-pared to RFs-surface wave inversion codes. Furthermore, the factthat only one data-derivative is used, makes this inversion codeextremely easy to use, without the need for complementarydatasets.
|
|
