International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

Pre-university education in general is facing various challenges, but in the schools of the Municipality of Kamenica the challenges are even greater in improving learning outcomes in teaching and learning, these challenges range from identifying individual student needs and optimizing learning resources.
This paper examines the role and impact of artificial intelligence (AI) in improving and facilitating learning circumstances in schools in the Municipality of Kamenica in Kosovo. The main goal is to analyze how AI solutions can improve the level of learning, support effective teaching and help in assessing and attracting students by increasing their level of concentration.
In this paper, we have taken an empirical approach using real data from several pre-university education schools in Kamenica, which include test scores, attendance statistics, and information on learning activities. With the help of machine learning techniques, models have been built that identify student profiles with different performance and recommend personalized learning strategies for each profile. An AIbased recommender system has also been developed that suggests teaching materials and relevant exercises using virtual laboratories according to the needs and individual progress of students.
The results show that the use of AI tools helps in identifying student weaknesses more quickly, in creating personalized lesson plans and in facilitating the work of teachers for continuous monitoring and evaluation. The analysis also shows the perception of teachers and parents across the Municipality of Kamenica towards the integration of AI in teaching practice, highlighting the challenges and opportunities for wider implementation in pre-university education in Kosovo.
This case study that we have conducted in the Municipality of Kamenica offers a practical and strategic contribution to municipal education policies towards the effective use of advanced technologies in the teaching process not only in the Municipality of Kamenica but also beyond.

Nonlinear dynamical systems arise in numerous scientific and engineering domains, including physics, economics, biology, and control theory. Their complex behavior, sensitivity to initial conditions, and possible chaotic dynamics make accurate modeling and prediction challenging using traditional analytical approaches alone. In recent years, artificial intelligence (AI) techniques have demonstrated strong potential for modeling nonlinear and complex systems through data-driven methods. This paper explores artificial intelligence approaches for modeling nonlinear dynamical systems, focusing on the integration of machine learning techniques with classical mathematical modeling. We consider representative nonlinear systems and analyze how neural networks, regression models, and hybrid AI–mathematical frameworks can be used to approximate system behavior, predict future states, and capture hidden structures in time-series data. Special attention is given to systems exhibiting chaotic behavior, where small perturbations in initial conditions can lead to significant divergence in trajectories. The study presents numerical simulations and comparative analyses between traditional mathematical models and AI-based approaches. The results highlight the advantages of machine learning methods in capturing nonlinear patterns and improving predictive accuracy, especially when analytical solutions are difficult or unavailable. Additionally, we discuss the interpretability of AI models in the context of dynamical systems and outline potential applications in engineering, intelligent control, and data-driven system identification. The proposed framework contributes to the growing intersection between dynamical systems theory and artificial intelligence by demonstrating how AI tools can support the analysis and modeling of complex nonlinear phenomena. This work aims to provide a foundation for future research on hybrid mathematical–AI methods for understanding and predicting complex systems.

High-voltage electrical discharge (HVED) treatment of powder mixtures is a modern, efficient, and economically advantageous method for both particle size reduction and modification of the material’s phase composition. The primary mechanisms of particle destruction within the discharge zone include shock waves, microcavitation, ablation, collisions with chamber components, and mutual abrasion between particles.
The application of machine learning methods to model HVED processes for titanium—a promising material for composite applications— enables more accurate predictions and optimization of the technological workflow.
The data used for modeling were obtained between 2013 and 2021 and include results from the treatment of the initial titanium powder (with an average diameter of d₀ = 60 μm) in ethanol. This setup enabled the formation of a volume-distributed multi-spark discharge (VMD) within the ethanol–powder dispersed system. The dataset includes information on the number of treatment pulses, discharge gap, pressure in the discharge channels, pressure on the chamber walls, and the amount of titanium carbide formed during the treatment process.
It was shown that the concentration of TiC gradually rises with the increase of specific treatment energy, regardless of the interelectrode gap. Specifically, at a specific energy (Ws) of 5 to 15 MJ/kg, the amount of titanium carbide reaches 10%; at 15 to 30 MJ/kg, it increases to 20%; and at energy levels above 30 MJ/kg, the TiC content reaches 30%.
Keywords: Ethanol, Titanium, Titanium Carbide, High-Voltage Electrical Discharge, Volume-Distributed Multispark Discharge, Electric Discharge Dispersion, Plasma Technologies, Machine Learning, Logistic Regression, Random Forest

FEM Simulation is widely used in engineering practice. The use in small and medium-sized companies is partially limited due to the high workload and the time required for the simulation calculations. In recent years, the use of Artificial Intelligence (AI) has been increasingly adopted, emerging as an exciting and promising area of research. This article presents a methodology for the implementation of artificial intelligence in the simulation process of a part and an assembly. This methodology includes phases to integrate AI into the CAD model preparation process, as well as the definition of contact conditions, fixtured reactions, and external forces. Artificial intelligence can process a large volume of previous calculations, allowing it to analyse and automate these preparation steps AND thus increase the accuracy of simulations.

Aluminum alloys are critical in industries such as aerospace and automotive due to their lightweight, strength, and corrosion resistance. Optimizing their properties is challenging and benefits from advanced predictive tools. This paper explores the use of mathematical modeling in understanding and designing aluminum alloys. Techniques like thermodynamic modeling (e.g., CALPHAD), phase transformation kinetics, and mechanical property simulations are reviewed. Computational methods, including finite element analysis and machine learning, are highlighted for their roles in alloy design and manufacturing, such as casting and additive manufacturing. Comparisons between model predictions and experimental results demonstrate accuracy and limitations. Applications in optimizing material properties and improving manufacturing processes are discussed. By accelerating alloy development and enabling tailored properties, mathematical modeling emerges as a transformative tool, advancing aluminum alloy research and driving innovation across industries.

OCR technology is widely used in various applications, including document digitization, data extraction, and document management systems. The OCR technology has seen significant advancements in recent years, especially with the integration of machine learning and artificial intelligence techniques. These advancements have led to substantial improvements in accuracy, particularly for standard fonts and clear document images. However, challenges still exist, especially when dealing with low-quality images, noisy images, handwritten text, or documents with unusual fonts. Some documents like ID cards, driving licenses, etc. use some security features like deliberate errors, OVI (Optical Variable Ink), Rainbow print, Guilloche pattern, fine line, and microprint to protect the documents from being counterfeited. These security elements also generate noise on the image to perform the OCR. In this paper, we present a way to enhance the OCR accuracy for ID-1 documents with security features through machine learning-driven image optimization. Albanian driving license images on the personalization process are used as a dataset to train the model. During the training process, the model of the ID-1 card is presented with a dataset containing input features (such as images, texts, or numerical data) along with corresponding labels or outcomes. After training, the model and the implemented algorithm to optimize the image for the OCR process are implemented in real-life application.

This article considers non-parametric models based on feature space modeling in the context of machine learning. The main machine learning models, their advantages and disadvantages are analyzed. The term “non-parametric feature space modeling model” has been considered in detail and compared with other machine learning models. The advantages of these models are justified in comparison with other approaches. The paper contains an analysis that confirms the advantages of using non-parametric feature space modeling models in machine learning tasks.

The goal of the paper is estimating the normalized friction torque of a joint in an industrial robotic manipulator. For this purpose a source data, given as a figure, is digitized using a tool WebPlotDigitizer in order to obtain numeric data. The numeric data is the used within the machine learning algorithm genetic programming (GP), which performs the symbolic regression in order to obtain the equation that regresses the dataset in question. The obtained model shows a coefficient of determination equal to 0.87, which indicates that the model in question may be used for the wide approximation of the normalized friction torque using the torque load, operating temperature and joint velocity as inputs.

Topological Data Analysis (TDA) is relatively new filed of Applied Mathematics that emerged rapidly last years. The main tool of Topological Data Analysis is Persistent Homology. Persistent Homology provides some topological characteristics of the datasets. In this paper we will discuss classification of digital images using their topological signatures computed with Persistent Homology. We will experiment on the Fashion-MNIST dataset. Using Topological Data Analysis, the classification was improved.

The ability to forecast the annual crop production is of crucial benefit for any country by providing the capability to define their import and export policies, as well as to estimate the economic gain of their agriculture planning. The weather conditions during the year significantly influence the growth of the crop, and the crop yield quantity is highly affected by the climate conditions in the different development cycles of the plant. Recently, the availability of historical climate data benefits the studies in the sector of agricultural sciences and food, and in particular the use of Artificial Intelligence methods in the big data analysis offers a significant opportunity to provide practicable information and actions. The present work aims to develop Machine Learning (ML) model to forecast the wheat yield based on historical climate data in a specific time frame in the Pelagonia valley in North Macedonia, as one of the most important regions for wheat production in the country. After pre-processing and selecting the input features, LS Boost regression model was employed as a ML method for estimation of the wheat yield from climate data, which resulted in high accuracy of wheat yield prediction even with limited dataset, both on the training and on the testing dataset. The research study proved the feasibility of using ML methods to complement the existing models for accurate wheat yield forecasting, providing significant advantage due to the ease of calibrating the ML model parameters.

The possibility of using machine learning methods to predict the results of high-voltage electric discharge treatment of titanium powder in a hydrocarbon liquid is studied. As a result of the work, distribution surfaces for the average particle diameter of Titanium powder. the amount of Titanium carbide formed during processing, and the number of spherical particles of titanium powder depending on the interelectrode gap and the number of pulses, when using spark discharge and with Titanium powder concentration in kerosene of 0.07 kg / dm3, pulse repetition frequency 0.3 Hz and the energy of single discharge of 1 kJ, were obtained.

Automatic classification of seismic sources finds vast application security domain. Due to restricted computational power in the edge devices a robust and less complex features need to be extracted and sent to appropriate classifier. Such features are histograms of oriented gradients. It represented the relative spectral distribution of derivative of amplitude against frequency instead of average spectral envelope.