Data science is the discipline that allows the exploration and analysis of data in order to extract useful and relevant information for decision making and problem solving. In the educational domain, human experiences need to be synthesized in order to improve the success rate and help the responsible to make the best informed decision. Analytic Hierarchical Process (AHP) is one of the most widely used multi-criteria analysis techniques in decision making. It allows building models for various problems even in the case of insufficient observation data. This paper aims to, benefit from the potentials of AHP technique, to analyze students’ profiles. Our objective is to detect and classify the most important fac-tors that increase Moroccan student dropout and failure. We expect that this study is the first one that explores AHP, studying the Moroccan context and describing student profiles depending on variant criteria. It reveals, on the one hand, that Moroccan student failure is strongly related to their family and behavioral characteristics. Indeed, lack of motivation, family instability and lack of responsibility are the top three factors causing failure at the university. On the other hand, student dropout is strongly related to studying context, namely the lack of orientation and repeated failures in modules. These findings will enable the decision makers to develop adequate solutions to overcome these two scourges.

Developments in information technology have led to the emergence of several online platforms for educational purposes, such as e-learning platforms, e-recommendation systems, e-recruitment system, etc. These systems exploit advances in Machine Learning to provide services tailored to the needs and profile of students. In this paper, we propose a state of art on student profile modeling using machine learning techniques during last four years. We aim to analyze the most used and most efficient machine learning techniques in both online and face-to-face education context, for different objectives such as failure, dropout, orientation, academic performance, etc. and also analyze the dominant features used for each objective in order to achieve a global view of the student profile model. Decision Tree is the most used and the most efficient by most of research studies. And academic, personal identity and online behavior are the top characteristics used for the student profile. To strengthen the survey results, an experiment was carried out, based on the application of machine learning techniques extracted from the state of art analysis, on the same datasets. Decision tree gave the highest performance, which confirms the survey results.

Student profile modeling is a topic that continues to attract the interest of both academics and researchers because of its crucial role in the development of predictive or decision support systems. It provides platforms to build intelligent systems such as e-orientation, e-recruitment, recommendation, and prediction systems. The purpose of this research is to propose an ontology-based decision support system that can be used for multi-objective prediction tasks such as prediction of failure/abundance, orientation or decision-making. Two major contributions are proposed here: a new domain ontology that models the profile of a student and a system that is based on this ontology to perform multiple prediction tasks. The proposed approach relies on the efficiency of the ontology to ensure semantic interoperability and the benefits of machine learning techniques to build an intelligent system for a multipurpose decision support objectives. The proposed system uses Decision Tree algorithm (C5.0), but other machine learning models can be added if they prove to be more efficient. Furthermore, the performance of the developed method is computed using performance metrics and achieved 83.6% for accuracy and 81.9% for recall.

The student profile has become an important component of education systems. Many systems objectives, as e-recommendation, e-orientation, e-recruitment and dropout prediction are essentially based on the profile for decision support. Machine learning plays an important role in this context and several studies have been carried out either for classification, prediction or clustering purpose. In this paper, the authors present a comparative study between different boosting algorithms which have been used successfully in many fields and for many purposes. In addition, the authors applied feature selection methods Fisher Score, Information Gain combined with Recursive Feature Elimination to enhance the preprocessing task and models’ performances. Using multi-label dataset predict the class of the student performance in mathematics, this article results show that the Light Gradient Boosting Machine (LightGBM) algorithm achieved the best performance when using Information gain with Recursive Feature Elimination method compared to the other boosting algorithms.

No abstract