Lejla Abazi Bexheti
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Lejla Abazi Bexheti has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Lediana Shala Riza
South East European University
3 shared publications - Arbana Kadriu
Faculty of Contemporary Sciences and Technologies , South East European University , Tetovo , North Macedonia
2 shared publications - Erind Bedalli
University of Elbasan
2 shared publications - Luan Sinanaj
Faculty of Contemporary Sciences and Technologies , South East European University , North Macedonia
2 shared publications - Luan Sinanaj
South East European University, North Macedonia
2 shared publications - Adrian Besimi
South East European University
1 shared publication
Latest Publications
- Open Access
- 25 September 2026
- Published by Walter de Gruyter GmbH in Triple Helix
- p. 1-28
University–Industry Collaboration and Innovation Performance: Insights from the Quadruple Helix Perspective
This paper examines the significance of university-industry collaboration under the Quadruple Helix approach in the development of innovation and productivity. It uses the 2023 dataset of the World Bank and EBRD BEEPS , covering 18,932 firms across 27 European economies. The analysis shows that university-industry collaboration is an important driver of innovation, and involvement in networks and civil society reinforces this impact. Probit, OLS, PPML and the IV methods are employed to investigate the university-industry collaboration and performance relationship. Research and development expenditure is found to be the most important driver of innovation, indicating the value of university collaboration for firms with high absorptive capacity. The paper contributes to the literature on the Quadruple Helix model and the talent ecosystem approach by integrating these models in the multi-country context. The analysis offers implications for managers and policymakers. It shows the value of university collaboration and R&D as drivers of innovation and the significance of stable environments as the foundation for the productive use of innovation incentives for regional growth.
- 25 September 2026
- Published by Walter de Gruyter GmbH in Triple Helix
- p. 1-28
Open AccessUniversity–Industry Collaboration and Innovation Performance: Insights from the Quadruple Helix Perspective
This paper examines the significance of university-industry collaboration under the Quadruple Helix approach in the development of innovation and productivity. It uses the 2023 dataset of the World Bank and EBRD BEEPS , covering 18,932 firms across 27 European economies. The analysis shows that university-industry collaboration is an important driver of innovation, and involvement in networks and civil society reinforces this impact. Probit, OLS, PPML and the IV methods are employed to investigate the university-industry collaboration and performance relationship. Research and development expenditure is found to be the most important driver of innovation, indicating the value of university collaboration for firms with high absorptive capacity. The paper contributes to the literature on the Quadruple Helix model and the talent ecosystem approach by integrating these models in the multi-country context. The analysis offers implications for managers and policymakers. It shows the value of university collaboration and R&D as drivers of innovation and the significance of stable environments as the foundation for the productive use of innovation incentives for regional growth.
- 2 January 2026
- Published by Springer Nature
- p. 687-699
Shifting from Quality Control to Quality Enhancement: Insights from QA-SURE Project Partner Institutions
- 2 January 2026
- Published by Springer Nature
- p. 687-699
Shifting from Quality Control to Quality Enhancement: Insights from QA-SURE Project Partner Institutions
- Open Access
- 1 December 2025
- Published by Walter de Gruyter GmbH in Business Systems Research Journal
- Vol. 16 (2) , 69-91
Early Identification of At-Risk Students in Online Education: A Deep Learning Approach to Predictive Modelling
Background: Predicting student performance in online learning is difficult due to class imbalance and limited model interpretability. At-risk students are fewer than high performers, biasing predictions, and methods like SMOTE fail to preserve temporal patterns. Although black-box models are accurate, they lack transparency for actionable insights. Objectives: This study proposes a deep learning framework combining LSTM networks and attention mechanisms to address these issues using the OULAD dataset. LSTMs capture temporal dependencies, while attention improves interpretability by emphasising key features. Advanced resampling mitigates class imbalance for robust at-risk student detection. Methods/Approach: The methodology applies the KDD framework to process data, uncover patterns, and build models that predict student success risk, ensuring efficient data handling, robust modelling, and actionable insights to improve retention. Results: The BiLSTM-RNN achieved the best performance, effectively capturing temporal dependencies and attaining the highest accuracy, precision, recall, and F1-score. Conclusions: The findings support more effective and targeted interventions in online education, offering valuable insights for research and practice.
- 1 December 2025
- Published by Walter de Gruyter GmbH in Business Systems Research Journal
- Vol. 16 (2) , 69-91
Open AccessEarly Identification of At-Risk Students in Online Education: A Deep Learning Approach to Predictive Modelling
Background: Predicting student performance in online learning is difficult due to class imbalance and limited model interpretability. At-risk students are fewer than high performers, biasing predictions, and methods like SMOTE fail to preserve temporal patterns. Although black-box models are accurate, they lack transparency for actionable insights. Objectives: This study proposes a deep learning framework combining LSTM networks and attention mechanisms to address these issues using the OULAD dataset. LSTMs capture temporal dependencies, while attention improves interpretability by emphasising key features. Advanced resampling mitigates class imbalance for robust at-risk student detection. Methods/Approach: The methodology applies the KDD framework to process data, uncover patterns, and build models that predict student success risk, ensuring efficient data handling, robust modelling, and actionable insights to improve retention. Results: The BiLSTM-RNN achieved the best performance, effectively capturing temporal dependencies and attaining the highest accuracy, precision, recall, and F1-score. Conclusions: The findings support more effective and targeted interventions in online education, offering valuable insights for research and practice.
- Open Access
- 3 October 2025
- Published by Learning Gate in Edelweiss Applied Science and Technology
- Vol. 9 (10) , 992-1004
Predicting road accident risks using web data: A classification approach
With the rapidly increasing rates of vehicle usage during recent decades, road accidents have become a significant concern, posing not only risks of injuries but also ranking among the leading causes of fatalities for young and middle-aged individuals. Several factors influence the occurrence of accidents, including careless driving, atmospheric conditions, speeding, and driving under the influence. Understanding the circumstances that lead to a greater risk of road accidents is very helpful for their prevention. The primary goal of this work is to explore patterns in road accidents that have occurred within the state of Albania based on web data scraped from news portals and reports from governmental institutions. The data mining pipeline first involves an intensive data preprocessing phase where scraping, filtering, and refining techniques are employed. Subsequently, several classification models are built on the preprocessed data. These models are developed using various methodologies, including naïve Bayes, random forests, XGBoost, and LightGBM. The constructed classification models are evaluated based on training-test splitting of the preprocessed data using various performance measures. Finally, these models can be used to predict the likelihood of accidents based on certain circumstances.
- 3 October 2025
- Published by Learning Gate in Edelweiss Applied Science and Technology
- Vol. 9 (10) , 992-1004
Open AccessPredicting road accident risks using web data: A classification approach
With the rapidly increasing rates of vehicle usage during recent decades, road accidents have become a significant concern, posing not only risks of injuries but also ranking among the leading causes of fatalities for young and middle-aged individuals. Several factors influence the occurrence of accidents, including careless driving, atmospheric conditions, speeding, and driving under the influence. Understanding the circumstances that lead to a greater risk of road accidents is very helpful for their prevention. The primary goal of this work is to explore patterns in road accidents that have occurred within the state of Albania based on web data scraped from news portals and reports from governmental institutions. The data mining pipeline first involves an intensive data preprocessing phase where scraping, filtering, and refining techniques are employed. Subsequently, several classification models are built on the preprocessed data. These models are developed using various methodologies, including naïve Bayes, random forests, XGBoost, and LightGBM. The constructed classification models are evaluated based on training-test splitting of the preprocessed data using various performance measures. Finally, these models can be used to predict the likelihood of accidents based on certain circumstances.
- Open Access
- 8 September 2025
- Published by International School for Social and Business Studies in International Journal of Management, Knowledge and Learning
- Vol. 14 (2)
Building HE Institutional QA system and QA offices: the QA-SURE Project in Kosovo and Albania
Through the Erasmus+ QA-SURE project, this study examines the institutional development of internal Quality Assurance (QA) offices and systems in higher education institutions (HEIs) in Kosovo and Albania. The initiative concentrated on creating or improving specialised QA offices at three HEIs in Albania and two in Kosovo, as opposed to more general national QA system changes. The study emphasises the ways in which training, participatory governance, strategic planning, and infrastructure support aided in the development of institutional capacity, drawing on ESG 2015. Results show advancements in digitisation, student and stakeholder interaction, and quality culture. The report makes the case that system alignment with European standards and sustained academic success are made possible by institution-driven QA systems.
- 8 September 2025
- Published by International School for Social and Business Studies in International Journal of Management, Knowledge and Learning
- Vol. 14 (2)
Open AccessBuilding HE Institutional QA system and QA offices: the QA-SURE Project in Kosovo and Albania
Through the Erasmus+ QA-SURE project, this study examines the institutional development of internal Quality Assurance (QA) offices and systems in higher education institutions (HEIs) in Kosovo and Albania. The initiative concentrated on creating or improving specialised QA offices at three HEIs in Albania and two in Kosovo, as opposed to more general national QA system changes. The study emphasises the ways in which training, participatory governance, strategic planning, and infrastructure support aided in the development of institutional capacity, drawing on ESG 2015. Results show advancements in digitisation, student and stakeholder interaction, and quality culture. The report makes the case that system alignment with European standards and sustained academic success are made possible by institution-driven QA systems.
- 3 August 2025
- Published by Springer Nature
- p. 3-14
Improving Situational Awareness During Natural Disasters Using XR Technologies
- 3 August 2025
- Published by Springer Nature
- p. 3-14
Improving Situational Awareness During Natural Disasters Using XR Technologies
- Open Access
- 24 April 2025
- Published by Association for Advancing Innovation and Research in Economy "Irenet" in ENTRENOVA - ENTerprise REsearch InNOVAtion
- Vol. 9 (1) , 60-71
A Classification Model for Predicting Road Accidents Using Web Data
The increase in urbanisation and the use of vehicles in recent decades has also led to increased road accidents. The causes of road accidents can be various, including human error, weather conditions or even inadequate road infrastructure. Knowing the causes and areas of road accidents can help prevent them by state institutions taking necessary measures and citizens being informed about the areas of road accidents. The primary purpose of this study is to explore patterns in accident web data in Albania and to construct a classification model using data mining techniques and methods. These techniques have been applied to data obtained from several leading media portals in Albania, including about 30,000 articles from online portals and reports from the state authorities. The constructed classification model is expected to be utilised to predict the accident likelihood according to the locations, weather, and period of the year.
- 24 April 2025
- Published by Association for Advancing Innovation and Research in Economy "Irenet" in ENTRENOVA - ENTerprise REsearch InNOVAtion
- Vol. 9 (1) , 60-71
Open AccessA Classification Model for Predicting Road Accidents Using Web Data
The increase in urbanisation and the use of vehicles in recent decades has also led to increased road accidents. The causes of road accidents can be various, including human error, weather conditions or even inadequate road infrastructure. Knowing the causes and areas of road accidents can help prevent them by state institutions taking necessary measures and citizens being informed about the areas of road accidents. The primary purpose of this study is to explore patterns in accident web data in Albania and to construct a classification model using data mining techniques and methods. These techniques have been applied to data obtained from several leading media portals in Albania, including about 30,000 articles from online portals and reports from the state authorities. The constructed classification model is expected to be utilised to predict the accident likelihood according to the locations, weather, and period of the year.
- 1 March 2025
- Published by Springer Nature
- p. 257-272
Predicting Student Engagement in Virtual Learning Environments Using ML Approaches with Data Balancing Techniques
- 1 March 2025
- Published by Springer Nature
- p. 257-272
Predicting Student Engagement in Virtual Learning Environments Using ML Approaches with Data Balancing Techniques
- Open Access
- 4 February 2025
- Published by Association for Advancing Innovation and Research in Economy "Irenet" in ENTRENOVA - ENTerprise REsearch InNOVAtion
- Vol. 10 (1) , 22-32
ML Techniques Integration in Digital Learning Platforms: Students’ Dataset Statistical Analysis
With the use of technology-enhanced learning platforms and an abundance of available educational data, it is possible to analyze student learning behavior and solve problems, improve the learning environment, and make data-driven decisions. A virtual learning environment effectively provide datasets for analyzing and reporting student learning, as well as its reflection and participation in their individual performances, which complements the learning analytics paradigm. This work is intended to explain the use of AI-based approaches in online learning, with a particular focus in offering a statistical approach on students VLE dataset. The study uses quantitative methodology to highlight the association between the variables in the obtained dataset. The purpose of this research is to examine the correlation and dependency of the dataset variables in order to observe the relationship between these variables and the effect that these attributes may have on students' performance in a digital learning environment. According to the findings of this study, there is a correlation between student performance and a number of different factors, such as resource (page) views, course modules, assessment type, assessment weight and sum of clicks in a VLE.
- 4 February 2025
- Published by Association for Advancing Innovation and Research in Economy "Irenet" in ENTRENOVA - ENTerprise REsearch InNOVAtion
- Vol. 10 (1) , 22-32
Open AccessML Techniques Integration in Digital Learning Platforms: Students’ Dataset Statistical Analysis
With the use of technology-enhanced learning platforms and an abundance of available educational data, it is possible to analyze student learning behavior and solve problems, improve the learning environment, and make data-driven decisions. A virtual learning environment effectively provide datasets for analyzing and reporting student learning, as well as its reflection and participation in their individual performances, which complements the learning analytics paradigm. This work is intended to explain the use of AI-based approaches in online learning, with a particular focus in offering a statistical approach on students VLE dataset. The study uses quantitative methodology to highlight the association between the variables in the obtained dataset. The purpose of this research is to examine the correlation and dependency of the dataset variables in order to observe the relationship between these variables and the effect that these attributes may have on students' performance in a digital learning environment. According to the findings of this study, there is a correlation between student performance and a number of different factors, such as resource (page) views, course modules, assessment type, assessment weight and sum of clicks in a VLE.
- Open Access
- 1 December 2023
- Published by Walter de Gruyter GmbH in SEEU Review
- Vol. 18 (2) , 176-190
Analysis and Visualization of Road Accidents Using Heatmaps Based on Web Data
Road accidents have increased rapidly in recent years for a variety of reasons. Analyzing and visualizing road accidents through heatmaps can help improve policies for their prevention by informing about areas with a high-risk of road accidents.The purpose of this research is to build a model for the analysis and visualization of road accidents through heatmaps. Information about road accidents is extracted from the news of the main online media portals through scripts in the Python language and Web Scraping techniques. From the extraction of about 30,000 articles from news portals for one year, only 829 were selected in the end that provided information about road accidents.As a result, and contribution of this research, a corpus was built with the geographic coordinates of road accidents and on this data our model was applied for the analysis and visualization of high-risk areas of road accidents using heatmaps. The visualization of heatmaps was done through a Python script, where it was applied to the geographic coordinates of road accidents.
- 1 December 2023
- Published by Walter de Gruyter GmbH in SEEU Review
- Vol. 18 (2) , 176-190
Open AccessAnalysis and Visualization of Road Accidents Using Heatmaps Based on Web Data
Road accidents have increased rapidly in recent years for a variety of reasons. Analyzing and visualizing road accidents through heatmaps can help improve policies for their prevention by informing about areas with a high-risk of road accidents.The purpose of this research is to build a model for the analysis and visualization of road accidents through heatmaps. Information about road accidents is extracted from the news of the main online media portals through scripts in the Python language and Web Scraping techniques. From the extraction of about 30,000 articles from news portals for one year, only 829 were selected in the end that provided information about road accidents.As a result, and contribution of this research, a corpus was built with the geographic coordinates of road accidents and on this data our model was applied for the analysis and visualization of high-risk areas of road accidents using heatmaps. The visualization of heatmaps was done through a Python script, where it was applied to the geographic coordinates of road accidents.