Exposing Social Data as Linked Data in Education

Abstract

According to recent studies, the social interactions of users such as sharing, rating, and reviewing can improve the value of digital learning objects and resources on the web. Linked data techniques, on the other hand, make different kinds of data available and reusable for other applications on the web. Exposing (meta)data, especially with a complex structure, as resource description framework (RDF) requires an ontology to bring all the data types under one umbrella. In this article, the authors propose an ontology in which social activities of users are exposed as linked data by reusing existing vocabularies. The proposed ontology has been implemented in a federated open educational resources (OER) portal, in which they published ratings, shares, comments, and other social activities assigned to around 1,000 OERs. This exposure allows other datasets, including harvested repositories, to explore the exposed social data related to e-learning objects according to the users' social engagement.

Convergence between emerging technologies and active methodologies in the university

“…In architecture and design, AR can overlay 3D models onto real-world environments, allowing students to visualize construction projects more accurately. Another essential application of emerging technologies is educational data analytics (Rajabi & Greller, 2019). Collecting and analyzing data on student performance and behavior provides educators with valuable insights.…”

Distributed and Deep Vertical Federated Learning with Big Data

“…Their approach aims at training a Logistic Regression (LR) model based on distributed data stored in a huge amount of global mobile users without conflicting local regulations or laws. Their approach is classified into horizontal FL years later, because of the wide range of participants and similar data features, e.g., health care with the attention mechanism on graphs 28 , 5G-empowered drone networks with reinforcement learning for smart grid or smart cities 29 , frequent itemset mining 30 , distributed medical data 31 and education data 32 , purchase behaviour with an attention mechanism 33 , and Internet of Things (IoT) devices 34,31 . In addition, the horizontal FL can be carried out in the Cloud for extreme gradient boosting 35 or deep learning models with the combination of synchronous 36,37,38,39,40 and asynchronous mechanisms 41,42 .…”

Distributed and deep vertical federated learning with big data

“…In 2016, the concept of FL was first introduced by McMahan et al 8 Their approach aims at training a logistic regression (LR) model based on distributed data stored in a huge amount of global mobile users without conflicting local regulations or laws. Their approach is classified into horizontal FL years later, because of the wide range of participants and similar data features, for example, health care with the attention mechanism on graphs, 28 5G‐empowered drone networks with reinforcement learning for smart grid or smart cities, 29 frequent itemset mining, 30 distributed medical data 31 and education data, 32 purchase behavior with an attention mechanism, 33 and Internet of Things (IoT) devices 31,34 . In addition, the horizontal FL can be carried out in the cloud for extreme gradient boosting 35 or deep learning models with the combination of synchronous 36‐40 and asynchronous mechanisms 41,42 .…”