Goran Keković

International Journal of Human–Computer Interaction

Exploring Learner Engagement in E-Learning Environments: A Predictive Analytics Perspective

Mikić, Vladimir; Keković, Goran; Mangaroska, Katerina; Ilić, Miloš; Kopanja, Lazar; Vesin, Boban

https://doi.org/10.1080/10447318.2025.2535530

During the last decade, the embracement of learner engagement in developing educational technologies has contributed to the amalgamation of favorable pedagogical practices and advanced learning tools. New opportunities for tailoring data-driven learning designs created optimal conditions for crafting personalized, interactive e-learning environments that foster successful learning outcomes. Although a plethora of metrics exist to capture engagement, there is a need for comprehensive research that incorporates both the learner’s subjective perceptions of their engagement and the objective indicators of their actual engagement. The goal of this research is twofold: first, to investigate the relationships between interaction data on student behavior in an e-learning environment and their self-reported engagement data, and second, to design a model for predicting students’ level of engagement based on the study findings. Statistical analysis was conducted using data from undergraduate students at the University of South-Eastern Norway who completed a one-semester programming course. Artificial neural networks were then used to develop a prediction model for classifying engagement levels. The findings highlight the importance of e-learning features such as coding exercises, topic-based assessments, and explanatory hints in fostering student engagement.

e-learning environment; learning engagement; behavior data; questionnaire

M21

Computer Science and Information Systems

Artificial Neural Network Modeling for Air Pollution Prediction: LSTM versus Levenberg-Marquardt Approach

Goran Keković; Rade Božović; Sonja Ketin; Vladimir Mikić; Miloš Ilić; Boban Vesin

In Press

Accurate prediction of air pollutant concentrations remains a critical challenge for environmental monitoring and public health, demanding robust and adaptive artificial intelligence approaches. This study investigates the effectiveness of various types of artificial neural networks, including Long Short-Term Memory networks (LSTM) and networks based on the Levenberg–Marquardt algorithm and its variant with Bayesian regularization (LMBR), in forecasting air pollution under different data conditions. The networks were tested on two datasets: the Air Quality dataset, with benzene concentration as the target variable, and the Beijing PM2.5 dataset, with PM2.5 concentration as the target. The efficiency parameters included MAE, RMSE and MAPE. The results show that LSTM and LMBR networks were competitive on the Air Quality dataset, while LSTM networks were more robust on the Beijing dataset, where relationships between PM2.5 concentration and predictor values were non-monotonic. The study also indicates that input variable selection can be used to detect seasonal trends in input data.

Artificial Intelligence; LSTM network; Levenberg–Marquardt algorithm; Input variable selection

M22

International Scientific and Professional Conference "ALFATECH – Smart Cities and Modern Technologies" 2026

Reframing Air Pollution Prediction: The Regression-Classification Dichotomy in Artificial Neural Networks

Goran Keković; Vladimir Mikić; Miloš Ilić; Rade Božović

https://conference.alfatech.rs/proceedings/

The problem of air pollution affects large urban areas and human health, while also threatening the environment, including flora and fauna, water, soil, ecosystems and climate change. Artificial neural networks are widely used in real-time air pollution prediction and forecasting. Although ANNs are commonly applied to regression tasks, they are often most effective in classification tasks. This paper examines the possibility of using Long Short-Term Memory (LSTM) neural networks as classifiers in real-time air pollution prediction tasks with PM2.5 particles, while also testing their properties in regression tasks. The results indicate that regression is effective for air pollution assessment, whereas classification is suitable as an alarm system in cases of air pollution hazardous to human life.

Air pollution; Classification method; LSTM networks; PM2.5 particles; Regression method

M31

Radioengineering

Artificial Bias Induction in Fourth-Order Cumulants Based Automatic Modulation Classification Algorithm in AWGN and Multipath Propagation Channel

Božović, Rade; Orlić, Vladimir; Keković, Goran

https://www.radioeng.cz/fulltexts/2025/25_02_0224_0233.pdf

Automatic modulation classification (AMC) is a widely used technique for recognizing modulation formats of signals considered a priori unknown. Due to low algorithm and hardware complexity, AMC algorithms based on fourth-order cumulants remain popular. This paper proposes a new AMC approach focused on manipulating theoretical expected cumulant values of real signal constellations through artificial bias induction. The method was evaluated through Monte Carlo simulations in AWGN and multipath propagation channels. Results confirmed the superiority of the proposed approach, especially at lower SNR values, with performance improvements up to 25%.

AMC; AWGN; bias; Binary Phase Shift Keying (BPSK); channel impulse response; cumulants; multipath; Quadrature Amplitude Modulation (QAM)

M23

Proceedings of International Scientific and Professional Conference “ALFATECH” Smart Cities and Modern Technologies

Optimization of AI Methods for Air Pollution Prediction

Keković, Goran; Božović, Rade

https://conference.alfatech.rs/

Air pollution is one of the major problems of large urban areas, and artificial intelligence methods can be used to predict pollution levels based on a wide range of parameters. This paper analyzes the use of artificial neural networks based on the Levenberg–Marquardt algorithm with Bayesian regularization. The results show high prediction accuracy and demonstrate that, in addition to tuned ANN parameters, optimal sample size can help balance desired accuracy and deviation between simulated and real data.

air pollution; artificial neural networks; Bayesian regularization; Levenberg–Marquardt algorithm

M33

IEEE Transactions on Learning Technologies

Predicting Student Performance in a Programming Tutoring System Using AI and Filtering Techniques

Ilić, M.; Keković, G.; Mikić, V.; Mangaroska, K.; Kopanja, L.; Vesin, B.

https://ieeexplore.ieee.org/document/10608149

This study examines the use of machine learning and artificial neural networks for predicting student grades in a programming tutoring system. The experiment used interaction data from university students and explored correlations among predictors and between predictors and the target variable. A filtering technique based on the minimum redundancy–maximum relevance criterion showed that data structure and correlations play a major role in defining an appropriate prediction model. ANN with the Levenberg–Marquardt algorithm and Bayesian regularization outperformed the tested ML methods.

artificial intelligence; accuracy; electronic learning; predictive models; artificial neural networks; statistical analysis; correlation

M21a

Proceedings of International Scientific and Professional Conference “ALFATECH” Smart Cities and Modern Technologies

A Modified Version of Gradient Descent Algorithm as a Solution of Local Minimum Problem in Artificial Neural Network

Keković, Goran; Božović, Rade; Stamenković, Negovan

https://conference.alfatech.rs/

This paper proposes software based on modified gradient descent and backpropagation algorithms for overcoming the local minimum problem in artificial neural networks. During network training, the existence of the global minimum was checked through successive loss function values and by determining the percentage of successfully classified samples from training and test sets. The software was written in C# in an object-oriented and modular manner.

artificial neural networks; neuron weights; global minimum; gradient descent; backpropagation algorithm; C# programming language

M33

Proceedings of International Scientific and Professional Conference “ALFATECH” Smart Cities and Modern Technologies

An Optimal Threshold Adaptation of Energy Detector in Cognitive Radio for Smart Cities Applications

Božović, Rade; Keković, Goran

https://conference.alfatech.rs/

Smart cities represent a modern technological concept of interconnection among different devices for collecting and processing data. Cognitive radio technology enables more efficient use of radio spectrum, while spectrum sensing is crucial for minimizing interference. Due to its low implementation complexity, a non-coherent energy detector is a suitable choice for spectrum sensing. This paper presents a closed-form solution for energy detector threshold adaptation in cognitive radio.

advanced smart communications; cognitive radio; energy detector; spectrum sensing; threshold adaptation

M33

Artificial Intelligence Conference – Book of Abstracts (Serbian Academy of Sciences and Arts, SASA)

Stability of Machine Learning Methods

Keković, Goran

https://www.mi.sanu.ac.rs/~ai_conf/previous_editions/2024/AI_Conference_Book_of_Abstracts.pdf

This work investigates the stability and generalization properties of machine learning methods for data classification using four different publicly available databases. The analyzed methods included support vector machine, ensemble of bagged trees, and k-nearest neighbors. The results showed that the bagged decision tree method achieved particularly strong performance, especially where statistical dependence between input variables was low or negligible.

support vector machines; ensemble of bagged trees; k-nearest neighbor; coefficient of variation; mutual entropy

M34

Acta Neurologica Belgica

Cognitive impairment and depression after acute myocardial infarction: associations with ejection fraction and demographic characteristics

Aleksandar Dikić; Ljiljana Radmilo; Željko Živanović; Goran Keković; Slobodan Sekulić; Zoran Kovačić; Ruža Radmilo

https://link.springer.com/article/10.1007/s13760-020-01440-0

Cognitive impairment and depression are often associated with acute myocardial infarction, although the risk factors for their occurrence after myocardial infarction are still unclear. This prospective study examined the effect of reduced ejection fraction and demographic characteristics on cognitive impairment and depression after myocardial infarction. The final sample consisted of 82 patients, divided according to ejection fraction values. The occurrence of cognitive impairment and depression was not significantly associated with ejection fraction, while demographic characteristics showed predictive effects. Age was a significant predictor of cognitive impairment, and female gender was a significant predictor of depression. The results indicate that cognitive impairment after acute myocardial infarction is more common in older patients, while depression is more common in women.

Cognitive impairment; Depression; Ejection fraction; Myocardial infarction

M22

Fetal and Pediatric Pathology

Inner Ear Malformations in Congenital Deafness Are Not Associated with Increased Risk of Breech Presentation

Sekulic, S.; Lemajic-Komazec, S.; Sokolovac, I.; Topalidou, A.; Gouni, O.; Petkovic, B.; Martac, Lj.; Kekovic, G.; Redžek-Mudrinic, T.; Capo, I.

https://doi.org/10.1080/15513815.2020.1737993

There is speculation that an immature vestibular system may be associated with breech presentation at delivery. The aim of this review was to determine whether syndromes with congenital inner ear malformations are accompanied by a higher frequency of breech presentation or malpresentation than in the general population. A literature review was conducted using PubMed/MEDLINE for cases of congenital deafness and vestibular malformations. A total of 122 cases were identified. The frequency of breech presentation was 1.64%, and transverse lie was also 1.64%, giving a total of 3.28% malpresentations. The results suggest that congenital malformations of the vestibular apparatus are not associated with an increased risk of breech presentation at delivery.

Congenital deafness; inner ear malformation; breech presentation; vestibular malformation; fetus; gestation

M23