Artificial Intelligence (AI) is increasingly used in online education platforms to provide valuable insights into students’ performance and success. However, the complexity of AI models makes it challenging for educators to interpret the specific factors that influence whether a student is going to pass or fail. Utilizing the Open University Learning Analytics Dataset (OULAD), this study employs various machine learning and deep learning techniques for predicting students’ success, along with SHapley Additive exPlanations (SHAP) as an Explainable Artificial Intelligence (XAI) technique, to understand the key factors behind success or failure. Unlike traditional statistical methods that explore variable relationships, this AI-driven approach uses advanced deep learning techniques to identify patterns and insights, allowing for a better understanding of the factors influencing student success. Additionally, this study focuses on identifying students at risk of failure using XAI techniques, specifically SHAP, to interpret model outputs by breaking down how specific factors contribute to a student’s success. This method enables targeted interventions to support their success. Results reveal that student engagement and registration timelines are critical factors affecting performance. The customized models achieve up to 94% accuracy for the designed tasks, outperforming traditional approaches. This study contributes to the use of AI in education and offers practical insights not only for educators but also for administrators and policymakers to enhance the quality and effectiveness of online learning.
The electricity load forecasting handles the process of determining how much electricity will be available at a given time while maintaining the balance and stability of the power grid. The accuracy of electricity load forecasting plays an important role in ensuring safe operation and improving the reliability of power systems and is a key component in the operational planning and efficient market. For many years, a conventional method has been used by using historical data as input parameters. With swift progress and improvement in technology, which shows more potential due to its accuracy, different methods can be applied depending on the identified model. To enhance the forecast of load, this paper introduces and proposes a framework developed on graph database technology to archive large amounts of data, which collects measured data from electrical substations in Pristina, Kosovo. The data includes electrical and weather parameters collected over a four-year timeframe. The proposed framework is designed to handle short-term load forecasting. Machine learning Linear Regression and deep learning Long Short-Term Memory algorithms are applied to multiple datasets and mean absolute error and root mean square error are calculated. The results show the promising performance and effectiveness of the proposed model, with high accuracy in load forecasting.
This study presents a comprehensive bibliometric analysis of research publications on artificial intelligence (AI) applications in higher education, with a particular focus on student performance-related studies. Drawing on 1431 documents retrieved from the Web of Science (WoS) Core Collections, advanced bibliometric tools, such as VOSViewer (1.6.19) and Biblioshiny, were used to explore research trends, citation networks, co-authorship patterns, and keyword co-occurrences. The results revealed an increase in AI-related educational research during the COVID-19 pandemic, reflecting the rise in the reliance on AI for enhancing learning outcomes in times of educational disruption. The findings also indicated that the research focus is gradually moving towards using AI for educational assessment, emphasizing the importance of accurate and data-driven evaluation of student performance. The co-occurrence of keywords and citation analyses confirmed that machine learning, deep learning, and predictive modeling are among the dominant AI techniques applied to assess and predict student outcomes. Furthermore, the study highlights AI’s potential in identifying learning gaps and enabling personalized interventions, allowing educators to address students’ specific needs more effectively. This new trend suggests a growing recognition of AI’s role in refining educational methodologies and improving performance evaluations at the tertiary level.
The evolution of electrical energy in all its aspects is one of the biggest challenges facing the operation of modern power systems. With the addition of numerous new, networked components to these systems, such as intelligent devices, electric vehicles, renewable energy sources, and battery storage, maintaining reliability and security is becoming increasingly difficult. To ensure that power transmission systems are secure and reliable, network security and reliability evaluations must be performed on a regular basis. The conventional approaches of using engineering tools for power flow analyses need to be enhanced in light of current grid challenges. More comprehensive analyses require the use of complex network concepts. This research proposes employing a combination of the Neplan engineering software and complex network concepts using Python NetworkX, both of which are recent advances, to address the challenges related to the modern grid by assessing the security and reliability of the power transmission system. The experimental results show enhanced reliability and security by employing complex network concepts to assess the grid’s topology and identify essential elements based on their centrality and betweenness and power flow analyses to comprehend how power flows and how various operating conditions impact the system.
22 December 2021 in Technology transfer: fundamental principles and innovative technical solutions
Big Data has been created from virtually everything around us at all times. Every digital media interaction generates data, from computer browsing and online retail to iTunes shopping and Facebook likes. This data is captured from multiple sources, with terrifying speed, volume and variety. But in order to extract substantial value from them, one must possess the optimal processing power, the appropriate analysis tools and, of course, the corresponding skills. The range of data collected by businesses today is almost unreal. According to IBM, more than 2.5 times four million data bytes generated per year, while the amount of data generated increases at such an astonishing rate that 90 % of it has been generated in just the last two years. Big Data have recently attracted substantial interest from both academics and practitioners. Big Data Analytics (BDA) is increasingly becoming a trending practice that many organizations are adopting with the purpose of constructing valuable information from BD. The analytics process, including the deployment and use of BDA tools, is seen by organizations as a tool to improve operational efficiency though it has strategic potential, drive new revenue streams and gain competitive advantages over business rivals. However, there are different types of analytic applications to consider. This paper presents a view of the BD challenges and methods to help to understand the significance of using the Big Data Technologies. This article based on a bibliographic review, on texts published in scientific journals, on relevant research dealing with the big data that have exploded in recent years, as they are increasingly linked to technology
22 June 2023 in International Journal of Membrane Science and Technology
Value investment is an attractive paradigm for individual investors. It involves different steps including evaluating past performance that could be challenging. We propose a representation for financial time series in a form appropriate for both human interpretation and automatic processing. We design a model for predicting sequence of values as opposed to point values. Combined with application of encoder-decoder type of neural network model architecture this allows interpretation of model parameters and intermediate activations by domain experts. We show that predictions better than the trivial last observed value are possible. Therefore, informed investment decisions can be supported by neural network models and the proposed representation and model interpretation.