Biljana Risteska Stojkoska
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Biljana Risteska Stojkoska has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
Danco Davcev
Faculty of Computer Science and Engineering, Intelligent Systems Department, Ss. Cyril and Methodius University, Rugjer Boshkovikj 16, 1000 Skopje, Macedonia
5 shared publicationsMarija Mihova
Faculty of Computer Science and Engineering, Ss Cyril and Methodius University, Skopje, Macedonia
3 shared publications- Alessandro Cantelli-Forti
CNIT
2 shared publications - Danco Davcev
Faculty of Electrical Engineering and Information Technologies University "Sts. Cyril and Methodious" Karpos 2, bb 1000 Skopje REPUBLIC OF MACEDONIA
2 shared publications - Edmond Jajaga
Mother Teresa University
2 shared publications - Eftim Zdravevski
Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University, Skopje, North Macedonia
2 shared publications
Latest Publications
- Open Access
- 14 April 2026
- Published by Springer Nature in Systematic Reviews
- Vol. 15 (1)
Optimizing document retrieval using massive text embeddings and LLM prompt engineering
Background: The rapid expansion of digital data poses a unique challenge for retrieving relevant and insightful information efficiently. In particular, the increasing volume of scientific publications has made literature reviews time-consuming. The emergence of large language models (LLMs) offers new opportunities to streamline this process. Methods: This paper explores the use of generative artificial intelligence (GenAI) for query reformulation and evaluates the performance of nine massive text embedding models, varying in size and fine-tuning strategies, in the context of document retrieval. We apply multiple prompt engineering techniques to evaluate the ability of LLMs to generate effective queries, comparing them with human-crafted queries. These are used to retrieve documents utilizing nine embedding models. The evaluation is across five datasets using metrics such as recall, average precision, and rank-based measures. Results: Results show that embedding models fine-tuned for semantic similarity consistently outperform general-purpose models, with UAE Large proving most robust across diverse domains. Furthermore, queries generated using zero-shot and few-shot prompting techniques often surpass the performance of human-formulated queries. Conclusion: These findings highlight the value of integrating LLMs and massive text embeddings to reduce manual effort in literature reviews. GenAI provides a reliable starting point for query formulation, with human input reserved for refinement when needed.
- 14 April 2026
- Published by Springer Nature in Systematic Reviews
- Vol. 15 (1)
Open AccessOptimizing document retrieval using massive text embeddings and LLM prompt engineering
Background: The rapid expansion of digital data poses a unique challenge for retrieving relevant and insightful information efficiently. In particular, the increasing volume of scientific publications has made literature reviews time-consuming. The emergence of large language models (LLMs) offers new opportunities to streamline this process. Methods: This paper explores the use of generative artificial intelligence (GenAI) for query reformulation and evaluates the performance of nine massive text embedding models, varying in size and fine-tuning strategies, in the context of document retrieval. We apply multiple prompt engineering techniques to evaluate the ability of LLMs to generate effective queries, comparing them with human-crafted queries. These are used to retrieve documents utilizing nine embedding models. The evaluation is across five datasets using metrics such as recall, average precision, and rank-based measures. Results: Results show that embedding models fine-tuned for semantic similarity consistently outperform general-purpose models, with UAE Large proving most robust across diverse domains. Furthermore, queries generated using zero-shot and few-shot prompting techniques often surpass the performance of human-formulated queries. Conclusion: These findings highlight the value of integrating LLMs and massive text embeddings to reduce manual effort in literature reviews. GenAI provides a reliable starting point for query formulation, with human input reserved for refinement when needed.
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 99501-99525
A Scenario-Based Evaluation Framework for Drone Detection: Analysis Under Distance, Altitude, and Operational Conditions
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 99501-99525
Open AccessA Scenario-Based Evaluation Framework for Drone Detection: Analysis Under Distance, Altitude, and Operational Conditions
- 23 April 2025
- Published by Springer Nature
- p. 136-146
Comparison of On-Board and Off-Board Processing Power Consumption for Drone Camera Images
- 23 April 2025
- Published by Springer Nature
- p. 136-146
Comparison of On-Board and Off-Board Processing Power Consumption for Drone Camera Images
- 2 August 2024
- Published by Springer Nature
- p. 177-191
Deep Dive into Invoice Intelligence: A Benchmark Study of Leading Models for Automated Invoice Data Extraction
- 2 August 2024
- Published by Springer Nature
- p. 177-191
Deep Dive into Invoice Intelligence: A Benchmark Study of Leading Models for Automated Invoice Data Extraction
- 4 August 2022
- Published by Springer Nature
- p. 422-433
An Image-Based Classification Module for Data Fusion Anti-drone System
- 4 August 2022
- Published by Springer Nature
- p. 422-433
An Image-Based Classification Module for Data Fusion Anti-drone System
- 1 January 2021
- Published by Centre for Evaluation in Education and Science (CEON/CEES) in Telfor Journal
- Vol. 13 (1) , 47-52
FINKI scholar, a publications database for Faculty of Computer Science and Engineering scholars
The aim of this paper is to develop a Web application where scholars of the Faculty of Computer Science and Engineering (FINKI) at the University of Ss. Cyril and Methodius can display and share their projects and publications. Visitors can view, search through, and filter the authors, projects and publications that can be added and edited by the administrators via the administrator panel. In this paper, we first explain the type of system we are building and go through similar existing systems explaining how they work and what they offer. Then, we go through the programming languages and technologies we decided to use to develop this Web application. After that, the development phase follows, where we describe each of the features we implemented. In The Final Product section we finally show images where you can see how the Web application works and what it looks like. We finish the paper with a conclusion, briefly summarizing what we have achieved.
- 1 January 2021
- Published by Centre for Evaluation in Education and Science (CEON/CEES) in Telfor Journal
- Vol. 13 (1) , 47-52
FINKI scholar, a publications database for Faculty of Computer Science and Engineering scholars
The aim of this paper is to develop a Web application where scholars of the Faculty of Computer Science and Engineering (FINKI) at the University of Ss. Cyril and Methodius can display and share their projects and publications. Visitors can view, search through, and filter the authors, projects and publications that can be added and edited by the administrators via the administrator panel. In this paper, we first explain the type of system we are building and go through similar existing systems explaining how they work and what they offer. Then, we go through the programming languages and technologies we decided to use to develop this Web application. After that, the development phase follows, where we describe each of the features we implemented. In The Final Product section we finally show images where you can see how the Web application works and what it looks like. We finish the paper with a conclusion, briefly summarizing what we have achieved.
- 14 October 2019
- Published by Springer Nature
- p. 41-51
Multidimensional Sensor Data Prediction
- 13 September 2018
- Published by Springer Nature
- p. 90-99
Framework for Human Activity Recognition on Smartphones and Smartwatches
- 13 September 2018
- Published by Springer Nature
- p. 90-99
Framework for Human Activity Recognition on Smartphones and Smartwatches
- 1 April 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 60, 1282-1289
Teaching computer programming: The macedonian case study of functional programming
The rapid development of information technology has imposed a high rate of change and reconceptualization in the design of computer science (CS) courses in order to expand access, provide flexibility in the learning environments, and meet the expanding needs of the ICT industry for qualified graduates. In such environment programming skills have become a core competence for computer science and engineering students. The issue of choosing and/or developing the most effective approach to teaching computer-programming languages has risen to the greatest importance. This paper presents a study on teaching functional programming using different teaching approaches combining different modes of delivering learning material and different assessment and feedback types. The analysis is done on data for a span of three years of teaching the Logical and Functional Programming (LFP) course and Artificial Intelligence (AI) course, both of which include functional programming during a 7 week, 4 hours per week, long period. We analyse what is the effect of different teaching approaches, but also what is the success of students taking both courses and which order of taking the exams improves the student results. Our results show that although moving toward online oriented teaching with high levels of automation is a necessity imposed by the rise in student numbers, we still need a dedicated face-to-face time and appropriate human feedback in solving programming problems. Additionally, the results imply that some reorganization in the Computer Science and Engineering Undergraduate Curricula is needed in order to meet student needs and preferences.
- 1 April 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 60, 1282-1289
Teaching computer programming: The macedonian case study of functional programming
The rapid development of information technology has imposed a high rate of change and reconceptualization in the design of computer science (CS) courses in order to expand access, provide flexibility in the learning environments, and meet the expanding needs of the ICT industry for qualified graduates. In such environment programming skills have become a core competence for computer science and engineering students. The issue of choosing and/or developing the most effective approach to teaching computer-programming languages has risen to the greatest importance. This paper presents a study on teaching functional programming using different teaching approaches combining different modes of delivering learning material and different assessment and feedback types. The analysis is done on data for a span of three years of teaching the Logical and Functional Programming (LFP) course and Artificial Intelligence (AI) course, both of which include functional programming during a 7 week, 4 hours per week, long period. We analyse what is the effect of different teaching approaches, but also what is the success of students taking both courses and which order of taking the exams improves the student results. Our results show that although moving toward online oriented teaching with high levels of automation is a necessity imposed by the rise in student numbers, we still need a dedicated face-to-face time and appropriate human feedback in solving programming problems. Additionally, the results imply that some reorganization in the Computer Science and Engineering Undergraduate Curricula is needed in order to meet student needs and preferences.
- Open Access
- 8 March 2017
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2017, 1-10
Internet of Things Framework for Home Care Systems
The increasing average age of the population in most industrialized countries imposes a necessity for developing advanced and practical services using state-of-the-art technologies, dedicated to personal living spaces. In this paper, we introduce a hierarchical distributed approach for home care systems based on a new paradigm known as Internet of Things (IoT). The proposed generic framework is supported by a three-level data management model composed of dew computing, fog computing, and cloud computing for efficient data flow in IoT based home care systems. We examine the proposed model through a real case scenario of an early fire detection system using a distributed fuzzy logic approach. The obtained results prove that such implementation of dew and fog computing provides high accuracy in fire detection IoT systems, while achieving minimum data latency.
- 8 March 2017
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2017, 1-10
Open AccessInternet of Things Framework for Home Care Systems
The increasing average age of the population in most industrialized countries imposes a necessity for developing advanced and practical services using state-of-the-art technologies, dedicated to personal living spaces. In this paper, we introduce a hierarchical distributed approach for home care systems based on a new paradigm known as Internet of Things (IoT). The proposed generic framework is supported by a three-level data management model composed of dew computing, fog computing, and cloud computing for efficient data flow in IoT based home care systems. We examine the proposed model through a real case scenario of an early fire detection system using a distributed fuzzy logic approach. The obtained results prove that such implementation of dew and fog computing provides high accuracy in fire detection IoT systems, while achieving minimum data latency.