This article focuses on the creation and evaluation of synthetic data to address the challenges of imbalanced datasets in machine learning (ML) applications, using fake news detection as a case study. We conducted a thorough literature review on generative adversarial networks (GANs) for tabular data, synthetic data generation methods, and synthetic data quality assessment. By augmenting a public news dataset with synthetic data generated by different GAN architectures, we demonstrate the potential of synthetic data to improve ML models’ performance in fake news detection. Our results show a significant improvement in classification performance, especially in the underrepresented class. We also modify and extend a data usage approach to evaluate the quality of synthetic data and investigate the relationship between synthetic data quality and data augmentation performance in classification tasks. We found a positive correlation between synthetic data quality and performance in the underrepresented class, highlighting the importance of high-quality synthetic data for effective data augmentation.

Practical programming competencies are critical to the success in computer science (CS) education and go-to-market of fresh graduates. Acquiring the required level of skills is a long journey of discovery, trial and error, and optimization seeking through a broad range of programming activities that learners must perform themselves. It is not reasonable to consider that teachers could evaluate all attempts that the average learner should develop multiplied by the number of students enrolled in a course, much less in a timely, deep, and fair fashion. Unsurprisingly, exploring the formal structure of programs to automate the assessment of certain features has long been a hot topic among CS education practitioners. Assessing a program is considerably more complex than asserting its functional correctness, as the proliferation of tools and techniques in the literature over the past decades indicates. Program efficiency, behavior, and readability, among many other features, assessed either statically or dynamically, are now also relevant for automatic evaluation. The outcome of an evaluation evolved from the primordial Boolean values to information about errors and tips on how to advance, possibly taking into account similar solutions. This work surveys the state of the art in the automated assessment of CS assignments, focusing on the supported types of exercises, security measures adopted, testing techniques used, type of feedback produced, and the information they offer the teacher to understand and optimize learning. A new era of automated assessment, capitalizing on static analysis techniques and containerization, has been identified. Furthermore, this review presents several other findings from the conducted review, discusses the current challenges of the field, and proposes some future research directions.

In recent years we have seen a large adherence to social media by various Higher Education Institutions (HEI) with the intent of reaching their target audiences and strengthen their brand recognition. It is important for organizations to discover the true audience-aggregating themes resulting from their communication strategies, as it provides institutions with the ability to monitor their organizational positioning and identify opportunities and threats. In this work we create an automatic system capable of identifying HEI Twitter communication strategies. We gathered and analyzed more than 18k Twitter publications from 12 of the top-HEI according to the 2019 Center for World University Rankings (CWUR). Results show that there are different strategies, and most of HEI had to adapt them to the covid situation. The analysis also shows the prediction of topics and retweets for a HEI cannot just be based on recent historical data.

The proliferation of false information on social networks is one of the hardest challenges in today's society, with implications capable of changing users perception on what is a fact or rumor.Due to its complexity, there has been an overwhelming number of contributions from the research community like the analysis of specific events where rumors are spread, analysis of the propagation of false content on the network, or machine learning algorithms to distinguish what is a fact and what is "fake news".In this paper, we identify and summarize some of the most prevalent works on the different categories studied.Finally, we also discuss the methods applied to deceive users and what are the next main challenges of this area.

Measuring the return on communication investments on social media has become one of the top key issues for organizations joining social networks. However, this field has been lacking articulation between what is conveyed as social media key performance indicators and the alignment of strategic organizational goals. Therefore, we propose a methodology to measure the performance of each organization on social media, to determine their positioning in the sector and to evaluate which are the content strategies used to boost the highest performing organizations. Thus, we identify how to determine which organizations should be closely monitored within the sector and which type content strategies can foster higher organizational performance on social media.

The expansion of social networks has contributed to the propagation of information relevant to general audiences. However, this is small percentage compared to all the data shared in such online platforms, which also includes private/personal information, simple chat messages and the recent called 'fake news'. In this paper, we make an exploratory analysis on two social networks to extract features that are indicators of relevant information in social network messages. Our goal is to build accurate machine learning models that are capable of detecting what is journalistically relevant. We conducted two experiments on CrowdFlower to build a solid ground truth for the models, by comparing the number of evaluations per post against the number of posts classified. The results show evidence that increasing the number of samples will result in a better performance on the relevancy classification task, even when relaxing in the number of evaluations per post. In addition, results show that there are significant correlations between the relevance of a post and its interest and whether is meaningfully for the majority of people. Finally, we achieve approximately 80% accuracy in the task of relevance detection using a small set of learning algorithms.

Predicting whether a student will pass or fail is one of the most important actions to take while giving lectures. Usually, the experienced teacher is able to detect problematic situations at early stages. However, this is only true for classes up to a hundred students. For bigger ones, automatic methods are needed. In this paper, we present a predictive system based on three criteria retrieved and computed from the logs of the learning management system. We built fast frugal decision trees to help predict and prevent student failures, using data retrieved from their resource usage patterns. Evaluation of the decision system shows that the system's accuracy is very high both in train and test phases, surpassing logistic regression and CART.

Social Media has been disrupting traditional technology mediated learning, providing students and educators with unsupervised and informal tools and spaces where authentic learning occurs. Still, the traditional LMS persists as the core element in this context, while lacking additional management, monitoring and analysis tools to handle informal learning and content. In this paper, we present an integrated methodology that combines social network analytics, sentiment analysis and topic categorization to perform social content visualizations and analysis aimed at integrated learning environments. Results provide insights on networked content dimension, type of structure, degree of popularity and degree of controversy, as well as on their educational and functional potential in the field of learning analytics.

The exponential growth of social media usage and the integration of digital natives in Higher Education Institutions (HEI) have been posing new challenges to both traditional and technology-mediated learning environments. Nowadays social media plays an important, if not central, role in society, for professional and personal purposes. However, it's important to highlight that in the mind of a digital native, social media is not just a tool, it is a place that is as real and as natural as any real-life world place where formal/informal social interactions happen. Still, formal higher education contexts are still mostly imprisoned in locked up institutional Learning Management Systems (LMS), while a new world of social connections grows and develops itself outside schools. One of the main reasons we believe to be persisting in the origin of the matter is the absence of a suitable management, monitoring and analysis tools to legitimize and to efficiently manage the relationship with students in social networks. In this paper we discuss the growing relevance of the "Social Student Relationship Management" concept and introduce the EduBridge Social system, which aims at connecting the most commonly used LMS, Moodle, and the most popular social network, Facebook.

In this paper we introduce three main features extracted from Moodle logs in order to be uses a possible means to predict future student grades. We discuss the statistical analysis on these features and show how they cannot be applied isolatedly to model our data. We then apply them as a whole and use principal component analysis to derive a decision tree based on the features. With derived tree we are able to predict grades in three intervals, namely to predict failures. Our proposed analysis methodology can be incorporated in an LMS and be used during a course. As the course unfolds, the system can to trigger alarms regarding possible failure situations.

The ability to handle large amounts of unstructured information, to optimize strategic business opportunities, and to identify fundamental lessons among competitors through benchmarking, are essential skills of every business sector. Currently, there are dozens of social media analytics' applications aiming at providing organizations with informed decision making tools. However, these applications rely on providing quantitative information, rather than qualitative information that is relevant and intelligible for managers. In order to address these aspects, we propose a semi-supervised learning procedure that discovers and compiles information taken from online social media, organizing it in a scheme that can be strategically relevant. We illustrate our procedure using a case study where we collected and analysed the social media discourse of 43 organizations operating on the Higher Public Polytechnic Education Sector. During the analysis we created an "editorial model" that characterizes the posts in the area. We describe in detail the training and the execution of an ensemble of classifying algorithms. In this study we focus on the techniques used to increase the accuracy and stability of the classifiers.

Group work is an essential activity during both graduate and undergraduate formation. Although there is a vast theoretical literature and numerous case studies about group work, we haven’t yet seen much development concerning the assessment of individual group participants. The problem relies on the difficulty to have the perception of each student’s contribution towards the whole work. We propose and describe a novel tool to manage and assess individual group. Using the collected interactions from the tool usage we create a model for predicting ill-conditioned interactions which generate alerts. We also describe a functionality to predict the final activity grading, based on the interaction patterns and on an automatic classification of these interactions.

Group work is an essential activity during both graduate and undergraduate formation. Students develop a set of skills, and employ criticism which helps them to better handle future interpersonal situations. There is a vast theoretical literature and numerous case studies about group work, but we haven't yet seen much development concerning the assessment of individual group participants. It is not always easy to have the perception of each student contribution to the whole work. Nevertheless, more than frequently, the assessment of the group is transposed to each group participant, which in turn results in each student having the same final mark. We propose and describe a tool to manage and assess individual group work taking into account the amount of work, interaction, quality, and the temporal evolution of each group participant. The module features the possibility to create two types of activities: collaborative or cooperative group work. We describe the conceptual design of our tool and present the two operating modes of the module, which is based on events, alerts and conditions. We then describe the methodology for the assessment in the two operating modes and how these two major approaches can be deployed through our module into pedagogical situations.

In this article we present a system capable of creating, managing and presenting digital portfolios. Our system innovates by using roles and states during its creation phase. This allows for high quality elements in the portfolio and promotes the students' reflection over them before full integration. The system also complies with the existing standards for e-portfolios. Moreover, it adds an extension to integrate previous created portfolios from different educational levels. In the article we show the need for such extension and describe how the system deals with integration of such diverse portfolios into a single one.

In this article, we detail a system that provides contributes for analyzing and characterizing interactions that occur between participants of online communities. We adapted and applied the Social Network Analysis methodology to online discussion forums to create a dynamical interaction graph. The graph can be embedded in learning managements systems and accessed through a web page. The functionality of the system provides a suitable environment to characterize the interactions between actors and their participations in discussion forums. In the article we describe the use of the system in two real-world situations. Our conclusions lead to the verification and the rapid identification of some important situations that occur in learning communities, such as: the location of actors more or less active; distinction of positions and roles; identification of different ways of organization/interaction in groups; characterization of the interactions of a group or of a community as a whole.

In this article we present a system capable of graphically representing the interactions between students and teachers in hierarchical online forums. By defining the a'reply-to' relation between the users the system builds a graph. During forum posts mining, the system computes metrics taken from social network analysis which are then applied to the graph drawing process. This system brings up new possibilities to e-learning as a tool capable of helping the teacher assorting and illustrating the degree of participation of students; to identify key students in information passing, and to find the implicit relations between forums participants. Preliminary tests lead to the conclusions that the system is able to rapidly help in identifying situations like outliers, sources and sinks of information. It also depicts rapidly sub communities formed from forum participants.

This paper describes a conceptual framework for implementing Intelligent Tutoring Systems using SCORM 2004. The main objective is to discuss how the SCORM 2004 sequencing and navigation specification can allow the development of Intelligent Web-Based Learning Environments using the sequencing and navigation tracking data, and rule set. Our main argument is that SCORM 2004 sequencing and navigation specification can be used to implement the two main functionalities of an ITS, (1) the inner loop and (2) the outer loop.

Learning objects are not a new concept, being currently produced either according to existent specifications, or in an ad hoc way. However, a system that can easily classify them for automatic storage is still missing. In this article we propose a repository of learning objects which is capable of automatic classification and categorization based on a semantic organization. The classification is based on a text extraction mechanism which then uses text mining methods. We present a methodology for the initial clustering of objects, and a technique for maintaining the organization with continuous arrival of new objects.

We propose a programming model for distributed concurrent systems with mobile objects in the context of a process calculus. Code mobility is induced by lexical scoping on names. Objects and messages migrate towards the site where their prefixes are lexically bound. Class definitions, on the other hand, are downloaded from the site where they are defined, and are instantiated locally upon arrival. We provide several programming examples to demonstrate the expressiveness of the model. Finally, based on this model we describe an architecture for a run-time system supporting concurrent, distributed computations and code mobility.