Search

At the beginning of the 2010 decade, the world of education and more specifically e-learning was revolutionized by the emergence of Massive Open Online Courses, better known by their acronym MOOC. Proposed more and more by universities and training centers around the world, MOOCs have become an undeniable asset for any student or person seeking to complete their initial training with free distance courses open to all areas. Despite the remarkable number of course enrollees, MOOCs have a huge dropout rate of up to 90%. This rate significantly affects the efforts made by the moderators for the success of this pedagogical model and negatively influences the learners’ experience and their supervision. To address this problem and help instructors streamline their interventions, we present a solution to classify MOOC learners into three distinct classes. The approach proposed in this paper is based on the filters methods to select the most relevant attributes and ensembling methods of machine learning algorithms. This approach has been validated by four MOOC courses from Stanford University. In order to prove the performance of the model (92.2%), a comparative study between the proposed model and other algorithms was made on several performance measures.

The Massive Open Online Course (MOOC) presents an approach of learning to permit an online and a distant learning for internet users. These systems are confronted with several problems among which the most important: the lack of participants’ interactivity in the platform and the dropout of the participants. Those challenges need to be resolved to ensure better engagement and to encourage the participants to complete the training, by improving their understanding-level. The present paper focuses on the proposition of a solution that helps to solve the lack of the participants’ interaction with the platform and to decrease the dropout rate. To do so, a module of virtual manipulations’ creation (virtual simulations and practical activities) is proposed. This last is based on Augmented Reality (AR) or Virtual Reality (VR), with a multitude of manipulation methods depending on the instructor choice. Two manipulation modes are proposed: the online mode, where all the manipulations and the interactions are done online and streamingly, and the offline mode, where the users can download the manipulations to preform them locally. Finally, to test the effectiveness of our proposition, two MOOCs, with the same thematic, have been proposed. The first one was a traditional MOOC while the second one was a MOOC that proposes to the participants some virtual simulations and practical activities.

Purpose Even if MOOCs (massive open online courses) are becoming a trend in distance learning, they suffer from a very high rate of learners’ dropout, and as a result, on average, only 10 per cent of enrolled learners manage to obtain their certificates of achievement. This paper aims to give tutors a clearer vision for an effective and personalized intervention as a solution to “retain” each type of learner at risk of dropping out. Design/methodology/approach This paper presents a methodology to provide predictions on learners’ behaviors. This work, which uses a Stanford data set, was divided into several phases, namely, a data extraction, an exploratory study and then a multivariate analysis to reduce dimensionality and to extract the most relevant features. The second step was the comparison between five machine learning algorithms. Finally, the authors used the principle of association rules to extract similarities between the behaviors of learners who dropped out from the MOOC. Findings The results of this work have given that deep learning ensures the best predictions in terms of accuracy, which is an average of 95.8 per cent, and is comparable to other measures such as precision, AUC, Recall and F1 score. Originality/value Many research studies have tried to tackle the MOOC dropout problem by proposing different dropout predictive models. In the same context, comes the present proposal with which the authors have tried to predict not only learners at a risk of dropping out of the MOOCs but also those who will succeed or fail.

As social networking has spread, people started sharing their personal opinions and thoughts widely via these online platforms. The resulting vast valuable data represent a rich source for companies to deduct their products’ reputation from both social media and crowds’ judgments. To exploit this wealth of data, a framework was proposed to collect opinions and rating scores respectively from social media and crowdsourcing platform to perform sentiment analysis, provide insights about a product and give consumers’ tendencies. During the analysis process, a consumer category (strict) is excluded from the process of reaching a majority consensus. To overcome this, a fuzzy clustering is used to compute consumers’ credibility. The key novelty of our approach is the new layer of validity check using a crowdsourcing component that ensures that the results obtained from social media are supported by opinions extracted directly from real-life consumers. Finally, experiments are carried out to validate this model (Twitter and Facebook were used as data sources). The obtained results show that this approach is more efficient and accurate than existing solutions thanks to our two-layer validity check design.

Since they were first set up in 2008, MOOCs have continued to integrate very deeply into the distance learning field. They have been adopted by a very large number of universities in order to complement face-to-face learning and thus remedy the massive number of students that the infrastructures can no longer support. In spite of the investments made for their development, MOOCs suffer from a huge drop-out rate of around 90%. This problem creates a number of difficulties for the instructors, including monitoring learners and group formation. In order to help the instructors to identify learners at risk of dropping out, this paper presents a model based on Multi-Layer Perceptron (MLP) that provides weekly predictions of each learner's engagement based on their behaviour. Our model has been tested on a data set of 3585 learners and has shown a high ability to identify this type of learner with an average accuracy of 90.3%.

E-learning systems suffer from a lack of tools to provide practical activities for learners. The mixed reality is promised to be a new technology to create a virtual environment where the learner is an actor by interaction with virtual objects. The objective is to establish a virtual laboratory that all tools and products can be manipulated by learners like in real practical work. This virtualization can also solve the safety problems and may reduce the risk of some experiments: nuclear, chemical, etc. Another interesting benefit is the reduction of the investment on real hardware (locally or remotely). We propose an approach for developing integrated E-learning systems, helping to carry out the practical work by the distance learner based on an augmented reality system.