This paper presents a comparative analysis of machine learning (SVM), deep learning (LSTM), and transformer-based (BERT) models for sentiment classification in Uzbek texts, enhanced by Named Entity Recognition (NER). The study addresses the challenge of accurately detecting sentiment in morphologically complex languages with limited resources, focusing on Uzbek–a Turkic language with rich agglutinative structures. A dataset of 10,000 user-generated comments from social platforms was annotated using a hybrid approach: manual labeling for sentiment (positive, negative, neutral) and a CRF-based NER system to identify entities (e.g., brands, locations, public figures). The integration of NER features aimed to resolve contextual ambiguities, such as distinguishing between "I love Samarkand’s history" (positive) and "Samarkand’s traffic is unbearable" (negative). Experimental results demonstrate that BERT, fine-tuned on Uzbek text, achieved the highest accuracy (90.2%) by leveraging contextualized embeddings to align entities with sentiment. LSTM showed competitive performance (85.1%) in sequential pattern learning but required extensive training data. SVM, while computationally efficient, lagged at 78.3% accuracy due to its inability to capture nuanced linguistic dependencies. The findings emphasize the critical role of NER in low-resource languages for disambiguating sentiment triggers and propose practical guidelines for deploying BERT in real-world applications, such as customer feedback analysis. Limitations, including data scarcity and computational costs, are discussed to inform future research on optimizing lightweight models for Uzbek NLP tasks.

This article presents a study on the creation of an emotional lexicon of the Uzbek language, taking into account its cultural and linguistic specificities. The emotional vocabulary of Uzbek reflects a complex interplay between language, national identity, and collective emotional experience. The purpose of the research is to develop a culturally sensitive framework for identifying and categorizing emotional expressions in Uzbek across diverse communicative contexts. The study analyzes a wide range of traditional sources, including folklore, proverbs, and classical literature, alongside modern discourse materials such as media texts, online forums, and social networks. Particular attention is given to dialectal variations and culturally embedded connotations, such as the double meaning of words like “achchiq” (bitter taste and emotional bitterness) and “suyanmoq” (physical leaning and emotional reliance). Methodologically, the research integrates corpus linguistics, sociolinguistic surveys, and AIbased sentiment analysis to ensure both empirical validity and cultural depth. As a result, the study proposes a prototype of an Uzbek emotional lexicon that captures emotional polarity, intensity, and contextual usage. The practical applications of this lexicon include improving human-computer interaction (e.g., chatbots), enriching language learning tools, and supporting sociolinguistic and affective computing research. The findings underscore the necessity of a standardized, culturally informed affective lexicon in Uzbek for preserving linguistic richness and enhancing emotional nuance in digital communication. This work contributes to the broader field of Turkic affective linguistics and emphasizes the importance of integrating cultural semantics into computational models.

The relevance of integrated data quality management tasks is increasing in the context of growing volume, variety, and criticality of data used. Despite this, organizations still have significant gaps in understanding the interconnections between data quality, process quality, and information systems. The purpose of this study is a systematic analysis of existing methodologies and concepts for data quality management, as well as the identification of key challenges in their practical implementation. The paper presents a review of scientific literature, outlines standard elements and a framework for an integrated approach to data quality. A high-level data flow scheme based on the Data Lakehouse architecture has been developed, reflecting the interaction of system components. The necessity of developing new methods and algorithms for optimizing big data quality, which go beyond traditional paradigms focused on structured data, is substantiated. Key problems often ignored in practice are identified and systematized, and criteria for the successful implementation of an integrated data quality management approach are formed.

The problem of studying the plane stress state by the photoelasticity method is considered. The technique is based on solving equilibrium equations. The boundary conditions for them are set based on the recorded interference pattern. A uniform grid is applied to it. For each node, the order of the interference band to which it belongs is determined. To date, the problem of automating this procedure has not been fully solved. To solve this problem, an algorithm has been developed and programmatically implemented that determines whether a node belongs to a stripe based on the color of the surrounding area. The algorithm is based on testing the statistical hypothesis that the sample belongs to a given distribution according to the Pearson criterion. To do this, the brightness histograms in all three color channels of the band of each order are quantitatively compared with the corresponding histograms constructed for the area in the vicinity of the node under consideration. The application of the method to the data taken at the PPU-7 installation has shown its effectiveness. In particular, the following results were obtained for interference patterns from simple objects (disk, plate). Of the 208 points (nodes of the rectangular grid) for which the interference band was determined, approximately 95% were correctly classified. Moreover, in some cases, there were no incorrectly classified pixels at all. Pixels that were not classified due to the fact that hypotheses about the correspondence of the chromaticity of their neighborhood to the color gamut of any of the bands were rejected accounted for 5-10%. 

This work is part of a series of articles reflecting the ATMO model (analysis of territorial multisectoral objects). Here, the previously described product and financial components of this model, implementing the Leontief input-output scheme, are complemented by the trade balance model. It comprehensively displays for each sector of the region the share of supplies of manufactured products and transit imports in a certain territorial system. These supplies are supported by existing product reserves. Further, these commodity flows are balanced by supplies to other regions outside this system and exports abroad. A special feature of the model is that it simultaneously takes into account the oncoming flows of domestic and foreign products into the modeled region from the cluster under consideration, while forming an import indicator into it. This value is used in the production balance block, which is also included in the ATMO model, thereby closing the complex of blocks included in it. In the trade balance block, commodity flows to other regions are determined by supply (transit) coefficients, according to which exports (imports) from a particular region are distributed into a cluster of regions for a particular sector. They are constructed as shares of the distribution of supplies of regional and foreign-made products across the regions of the corresponding cluster, based on a retrospective analysis of similar supplies in the past. Expressions are constructed to reflect the relationship between investments in transport infrastructure and the costs of inter-regional supplies of goods and equipment. The targets of this block determine the development of the regional system aimed at increasing exports of products produced in the regions. It, in turn, relies on the conditions of industrial growth provided by the ever-increasing demand for the supply of products outside each region and for export.

The article presents a study of errors arising in non-contact measurement of fuel assembly head height variations using the parallax-shift method under conditions of arc-shaped motion of a television camera. The primary sources of error are examined, including radiation-induced and noise-related distortions of the video stream, contour detection inaccuracies, subpixel deviations of circle centers, geometric imperfections of camera motion, and variations in camera orientation. A model describing the influence of these factors on the final height measurement accuracy is developed. The sensitivity of height estimations to key system parameters is analyzed. The study demonstrates that the method maintains metrological reliability under typical operational deviations characteristic of nuclear power plant environments.

As the complexity of information systems increases and the requirements for information security become more stringent, there arises a need not only for the collection and analysis of monitoring data but also for ensuring their immutability over time. Traditional monitoring systems do not provide protection against unauthorized modifications of metric history, which limits their applicability in contexts that require transparent auditing. This work proposes to address this issue using blockchain technology. The proposed approach involves daily extraction of key monitoring metrics for the previous day, followed by hashing of the aggregated data. The resulting hash is transmitted to a smart contract deployed on the public Ethereum blockchain, where it is stored along with a timestamp. For verification, a software module was developed that retrieves data from the monitoring system and compares their hash with the one stored on the blockchain. A prototype implementation has achieved full automation of the process of recording and subsequently verifying aggregated metrics. The verification procedure reliably detects any data tampering by comparing hashes. While the proposed method does not eliminate the risk of recording tampered data in the event of a trusted party being compromised, it does ensure a transparent and immutable history of recordings, allowing for retrospective detection of violations. A key advantage is its independence from the organization's internal infrastructure and the ability to perform verification using open tools.