Journal of Problems in Computer Science and Information Technologies

A COMPARATIVE STUDY OF LARGE LANGUAGE MODEL FOR CONCENTRATION PREDICTION OF OIL SLUDGE WITH NON-STATIONAL HEAT TRANSFER

2025-06-09T11:33:23+00:00

As data accumulates and computational power increase, the performance of large language models (LLMs) have been significantly improved, which promoted it entered a stage with rapid development in various research field. In order to explore the application capability of LLMs in complex physical problem, we took oil sludge as the research object to predicted the concentration based on the temperature at the corresponding location, using the dataset with dynamic viscosity uf equal 2.5 and 5.0 for training and testing. We selected six LLMs for experiments, and found that four of them had hallucination problems which was the outputs were inconsistent with the actual program. Then, we built a random forest (RF) and compared it to the RF model predicted by LLM in five-fold cross validation, to verify whether the parameters were potential optimized or not. Results shows that the artificial model was superior to the LLMs solution in terms of generalization and accuracy, whose RMSE and R2 are 0.02158, 0.9690.

SPATIOTEMPORAL ASSESSMENT OF SOIL SALINITY IN IRRIGATED AGRICULTURAL LANDS OF KAZAKHSTAN USING REMOTE SENSING

2025-05-06T17:16:45+00:00

Soil salinization poses a significant threat to agricultural productivity and environmental sustainability, particularly in arid and semi-arid regions. This study presents a comprehensive spatiotemporal analysis of soil salinity dynamics in irrigated lands of Alakol District, Zhetisu Region, Kazakhstan, using multi-temporal Sentinel-2 satellite imagery and the Normalized Difference Salinity Index (NDSI). The analysis covered the 2024 growing season, from March to November, with one cloud-free image selected for each month. NDSI values were calculated monthly and classified into four salinity categories: non-saline, slightly saline, moderately saline, and highly saline. Field sampling at 31 locations provided electrical conductivity (EC) data for validation, enabling comparison between surface reflectance-based salinity estimates and ground measurements. The results demonstrated pronounced seasonal trends: NDSI values were lowest in spring due to leaching by precipitation and early irrigation, gradually increasing through summer as evaporation concentrated salts at the surface, and fluctuating in autumn depending on rainfall and drainage conditions. Spatially, fields situated in topographic depressions or near Lake Alakol exhibited the highest salinity levels, whereas upland areas remained relatively unaffected. Notably, no fields exceeded the moderate salinity threshold, indicating that while salinization is present, it remains in early stages. The NDSI approach proved effective for surface salinity detection, capturing both temporal fluctuations and spatial heterogeneity. These findings underscore the utility of remote sensing for operational salinity monitoring and highlight the importance of continuous observation to inform timely land management interventions. This study offers actionable insights for sustainable agriculture, particularly in tailoring irrigation and drainage strategies to mitigate salinity risks across vulnerable farmlands in Central Asia.

MODELING OF THE DEFORMED STATE OF MESH PLATES USING COMPLEX CONFIGURATION

2025-06-04T05:55:15+00:00

In this article, the deformed state processes of mesh plates with complex configurations are modeled mathematically. Specifically, a computational algorithm comprising of the R-function methods of V.L. Rvachev (RFM) and the Bubnov-Galerkin method is applied. The mathematical model describes the behavior of mesh plates under external loads by representing equilibrium equations in a Cartesian coordinate system. The solution structures are built using constructive RFM approaches, and discretization is carried out with the Bubnov-Galerkin technique. Computational experiments are conducted to determine the deformation characteristics of mesh plates with intricate geometries. The proposed approach significantly reduces the computational complexity and increases the accuracy of results when compared to conventional analytical methods. Furthermore, the algorithm enables numerical analysis of rhomboidal and hexagonal plate configurations under different boundary conditions. These results may be utilized in designing lightweight yet structurally efficient components in aerospace, civil, and mechanical engineering.

PHASE-BASED INTERFEROMETRIC METHOD FOR PRECISE DISPLACEMENT ESTIMATION: THEORY AND COMPUTATIONAL POTENTIAL

2025-07-10T06:39:07+00:00

This paper presents the theoretical foundation and computational modeling of a novel phase-based interferometric method for precise displacement estimation in environmental monitoring applications. The method leverages the phase difference between two coherent radio signals transmitted over a wireless forward link, enabling sub-millimeter resolution without relying on reflected signals or embedded sensors. Unlike radar interferometry and distributed fiber optic systems, the proposed technique operates entirely in a forward-link architecture, making it more scalable, energy-efficient, and suitable for low-infrastructure deployments. Special attention is given to the computational procedures required for real-time signal interpretation, including instantaneous phase extraction using the Hilbert transform, phase unwrapping algorithms, and noise mitigation via digital filters. Simulation results confirm that the method is theoretically robust and computationally tractable, offering a practical path toward implementation using lightweight embedded platforms such as software-defined radios (SDRs) with GPS-disciplined oscillators. The results also demonstrate how design parameters such as carrier frequency and dual-tone spacing – affect the sensitivity and resolution of displacement estimates. This study lies at the intersection of applied computer science, signal processing, and geospatial engineering. It provides both a mathematical and algorithmic foundation for future systems aimed at distributed, real-time sensing in civil infrastructure and geohazard management.

OPTIMIZATION OF MARKETING STRATEGIES IN THE AGRO-INDUSTRIAL COMPLEX OF KAZAKHSTAN BASED ON A HYBRID METHOD

2025-08-22T09:59:09+00:00

In the development of the agro-industrial complex (AIC) of the Republic of Kazakhstan, one of the key aspects is the improvement of information and consulting activities of enterprises and companies in the agricultural sector, in particular, aimed at increasing the efficiency of production and marketing of agricultural products. A significant aspect is also the development of agromarketing in the AIC, which will ultimately improve market mechanisms and increase the competitiveness of products of local agricultural producers. The study proposes a hybrid method for solving the problem of multi-criteria optimization of marketing strategies in the AIC. The method combines the use of the NSGA-II algorithm and machine learning based on K-means to analyze the results. The quality of the solution was assessed using the hypervolume parameters and visualization of the found optimal strategies using the Pareto front. The theoretical and practical significance of the study is confirmed by the possibility of adapting the proposed hybrid method for enterprises of the AIC of Kazakhstan, taking into account existing regional restrictions.

APPLICATION OF RULE-BASED METHOD FOR AUTOMATIC EXTRACTION OF TAGS FROM COLUMN-STYLE PDF-DOCUMENTS

2025-09-24T10:09:30+00:00

This study presents a rule-based hybrid pipeline for the automated extraction of structured metadata from PDF versions of Kazakh-language newspaper articles, focusing on the national newspaper Egemen Qazaqstan. The primary goal is to support the development of a machine-readable knowledge base for future use in training large language models (LLMs) and building an AI-powered assistant for data journalism in Kazakhstan. The pipeline integrates three open-source parsers – pdfminer.six, PyMuPDF, and pdfplumber – to extract key elements such as title, author, date, abstract, text, journal name, and category. To evaluate extraction quality, we compared the results of the automated parser against manually annotated reference files across three real-world issues of the newspaper. The evaluation employed three complementary metrics: Precision, Textual Semantic Similarity (TSS), and Holistic Precision, which jointly assess both exact and semantic matches. The experimental results show that most tags – especially structured fields like date, journal, and category – achieved perfect Holistic Precision (1.00), while more variable fields like title still scored above 0.85. The validated pipeline was then applied to the full corpus of 2,140 newspaper PDFs published between 2017 and March 2025, successfully converting 159,135 articles into structured JSON format. This enriched corpus serves as a foundational knowledge base for Kazakh-language AI systems in journalism and media analysis.

DETECTION OF MENTAL DISORDERS BASED ON THE ANALYSIS OF EMOTION, FACIAL EXPRESSIONS AND FACIAL MOVEMENTS IN A VIDEO STREAM

2025-08-27T12:37:46+00:00

Traditional emotion recognition systems often rely on generalized person-centered models that do not consider the variability of individual emotion expression. This paper explores individual differences in emotion expression and facial expressions for recognizing mental disorders based on video streaming. Using machine learning techniques and deep learning algorithms, we aim to create an algorithm for emotion recognition using a personalized approach. The paper discusses the data collection methods, the condition and the impact of personalization on recognition accuracy. Experimental results demonstrate the advantages of automated personalized models over traditional models, highlighting their potential in the field of affective computing. The study also addresses ethical implications, advocating for bias-mitigated training through cross-cultural datasets and user-controlled calibration. With help of real-time edge computing, our system enables scalable, privacy-preserving mental health monitoring, underscoring the transformative potential of adaptive affective computing and remote diagnostics.

Forecast of Housing Prices in Almaty Using Machine Learning Algorithms

2025-08-26T05:15:24+00:00

Precise prediction of housing values is an important task for various stakeholders involved in the housing market, including investors, builders, and city planners. In this research, supervised machine learning models are used to predict the price of apartments in Almaty, Kazakhstan, which is a dynamic urban market in Central Asia. With an openly available dataset of apartments for sale, Linear Regression, Lasso Regression, Random Forest, and XGBoost models are implemented and tested. The data is scaled and encoded with scalable pipelines, and models are evaluated with regards to Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R² Score. The best performing model amongst those tested was Random Forest Regressor with an R² of 0.9158, followed by XGBoost with 0.8438. Feature importance visualization identifies district, area, and construction year as primary influencing factors. The research supports that ensembling machine learning models are efficient and scalable predictors for housing forecasts and suggests future improvements with time-series and geospatial features

WATER QUALITY MONITORING USING REFRACTIVE INDEX SENSING E-FBG SENSORS

2025-08-21T12:42:04+00:00

The need to protect the environment has stimulated the development of numerous analytical techniques for detecting pollutants in natural ecosystems, including methods for determining nitrate concentrations in source water. In this context, the present study introduces an experimental approach for water quality assessment based on etched fiber Bragg gratings (e-FBG). Specifically, the method relies on monitoring the shift in the Bragg wavelength, which occurs as a result of variations in the refractive index of water caused by changes in its chemical composition. Moreover, we proposed a water quality monitoring strategy employing e-FBG sensors, which provides high sensitivity to fluctuations in the optical properties of the surrounding medium. The applicability of the proposed sensor is demonstrated through the detection of low concentrations of nitrates in aquatic environments. The e-FBG sensor exhibits several notable advantages. In particular, it offers high resolution for wavelength shift detection, a high optical signal-to-noise ratio of 40 dB, and a narrow bandwidth of 0.02 nm, which collectively enhance the accuracy and reliability of peak wavelength measurements. Furthermore, the sensor supports optical remote sensing, making it suitable for real-time environmental monitoring. Therefore, the experimental results strongly suggest that the proposed e-FBG sensor holds significant potential for pollutant detection in practical field applications.

INFORMATION SYSTEM FOR METALLURGICAL PROCESS ANALYSIS AND OPTIMIZATION

2025-09-23T14:45:47+00:00

The metallurgical industry faces increasing challenges in reconciling production efficiency with environmental compliance while managing heterogeneous data streams across complex processing operations. Traditional approaches to metallurgical process analysis rely on manual calculations and isolated software tools, limiting operational efficiency and introducing potential errors in critical decision-making processes. This paper presents the design and implementation of a comprehensive web-based information system specifically developed for integrated metallurgical process analysis and optimization. The system architecture employs a modular design incorporating three specialized computational modules: pyrometallurgical calculations for ore-to-metal conversions, hydrometallurgical process modeling for extraction operations, and auxiliary process calculators for specialized applications. The platform integrates Django-based backend processing with responsive frontend interfaces, supporting multi-user access, comprehensive data validation, and seamless integration with existing plant information systems. Implementation includes predictive analytics capabilities utilizing machine learning algorithm for forward process prediction and optimization. System validation demonstrates robust performance with processing times ranging from 0.6 to 3.4 seconds across different computational modules and operational success rates exceeding 98.7% for all core functions. The platform supports multiple data input formats including manual entry and Excel file processing, with comprehensive export capabilities (JSON, CSV, Excel) enabling integration with downstream analysis tools. Performance evaluation indicates the system successfully addresses key industrial requirements for accuracy, reliability, and scalability in metallurgical process analysis applications. The developed architecture provides a practical framework for implementing digital transformation initiatives in metallurgical operations while maintaining computational precision required for critical industrial applications