This study aims to improve the accuracy, speed, and safety of suicide risk assessment among adolescents in the digital ecosystems of smart cities. To achieve this goal, an integrated system architecture was developed that combines natural language processing methods, transformer models, and privacy-preserving computation. The methodological part includes large-scale textual data analysis, distributed processing in Apache Spark and Hadoop environments, and the use of federated learning, which allows models to be trained without transferring sensitive source information. The evaluation was conducted on open mental health datasets and supplemented by a series of experiments simulating the system's operation in real time, as well as surveys of specialists – psychologists, educators, and IT experts. The analysis showed that transformer models, particularly BERT, significantly outperform classical algorithms, achieving an AUC-ROC of 0.96 and an F1 score of 0.92 with an average response time of 2.4 seconds. Survey participants noted the importance of transparency and data protection, and the proposed architecture received high marks for reducing the risk of information leaks and providing robust audit mechanisms. The novelty of the work lies in the combination of predictive analytics, federated learning, differential privacy, and blockchain traceability in a single application-oriented system. The results show that ethically sound and rapid suicide risk detection can be implemented in schools, medical institutions, and municipal services, providing both practical benefits and contributing to methodological advancements.

Detecting botnets in IoT environments is difficult because most intrusion detection systems treat network events as independent observations. In practice, infections spread through device relationships and evolve through distinct temporal phases. A system that ignores either aspect will miss important patterns. This paper explores a hybrid architecture combining Graph Neural Networks with Long Short-Term Memory networks to capture both structural and temporal dynamics. The GNN component models behavioral similarity between traffic flows in feature space, while the LSTM tracks how patterns change as attacks progress. The two components are trained jointly so that relational context is preserved during temporal learning. We evaluated the approach on two datasets with different characteristics. N-BaIoT contains traffic from nine devices infected with Mirai and BASHLITE, while CICIoT2023 covers 105 devices across 33 attack types. On N-BaIoT, the model achieved 99.88% accuracy with F1 of 0.9988 and Brier score of 0.0015. Cross-validation on CICIoT2023 yielded 99.73% accuracy with Brier score of 0.0030. The low Brier scores suggest that probability outputs are reasonably well calibrated for risk-based decision making. Consistent performance across both datasets provides some evidence that the architecture generalizes beyond a single benchmark setting.

This article presents a formal demonstration of a hybrid big data processing architecture that combines the fault tolerance and storage robustness of Hadoop with the speed and in-memory processing capabilities of Apache Spark. The proposed architecture is evaluated through test execution and performance benchmarking in real-world data centers across three regions in Kazakhstan. The model integrates distributed resource management components, Directed Acyclic Graph (DAG)-based scheduling mechanism, and Resilient Distributed Datasets (RDDs) to enable dynamic workload distribution and rapid failure recovery. The results demonstrate that the hybrid system consistently outperforms standalone Spark and Hadoop architectures under variable workloads, illustrating enhancements in execution time, task recovery, and resource utilization. Quantitative performance metrics allow for a structured comparison of architectures and help optimize deployments for diverse scenarios. The proposed hybrid architecture shows significant improvements, reducing average execution time by up to 38% and increasing resource efficiency by 25% compared to standalone Spark and Hadoop systems.

This article analyzes the current state of railway automation and telemechanics systems operating on the railway network of Kazakhstan. It identifies their critical indicators in terms of operational lifespan and reveals key issues within the mainline network. The division of station systems into zones of criticality based on operational lifespan is conducted considering the class of the station and the cargo intensity of maneuvering areas. The article proposes the development of a national microprocessor-based system for centralizing switches and signals as a solution to the identified existing problems. For this purpose, the authors have developed and proposed an operational algorithm for the control controller of the microprocessor-based centralization system using HIMA CENELEC SIL4 (controllers compliant with Safety Integrity Level 4 (SIL 4) according to CENELEC standards by HIMA) controllers to manage switches and signals at stations. The authors also present the architecture and key features of the developing microprocessor-based system. Comprehensive indicators of the functional safety of the national microprocessor centralization system with modular architecture (hereinafter - KZ-MPC-MA) are defined to demonstrate compliance with the SIL4 according to CENELEC (European Committee for Electrotechnical Standardization) and TR/CU (Technical regulations in force within the Customs Union) requirements.

The timely detection of psycho-emotional risks has become increasingly important due to the rapid growth of social media platforms. This study examines user-generated text as a potential source of early indicators of psychological vulnerability. The proposed NLP-based framework incorporates behavioral features to improve the interpretation of users’ psycho-emotional states. In addition to text classification, the study considers structured behavioral indicators to support psycho-emotional risk analysis. Particular attention is given to interpretability. SHAP-based techniques are applied to reveal the contribution of individual features and to provide a clearer explanation of model predictions. The evaluation was conducted on publicly available datasets containing textual data and aggregated behavioral/physiological indicators. No raw physiological streams, wearable sensor data, or biometric recordings were used. The two datasets were employed in complementary experimental settings and were not aligned at the individual-sample level; accordingly, the broader analytical perspective explored in this study should not be interpreted as a single end-to-end or fully aligned multimodal learning framework. The proposed BERT-based model with SHAP interpretability achieved an accuracy of 96.3%, an F1-score of 0.96, and a ROC–AUC score of 0.98, showing consistent improvement over baseline models, including Random Forests and Support Vector Machines.

This study presents a methodology for cross-ranking analysis to evaluate the consistency of universities’ positions across the QS World University Rankings, Times Higher Education (THE) World University Rankings, and Academic Ranking of World Universities (ARWU). Data for 2022–2025 were harmonized by university names, countries, years, and ranking positions, resulting in a unified dataset of 316 observations. Ranking agreement was assessed using the overlap of ranked lists, a Spearman distance–based similarity measure, and a normalized inverse-rank metric. QS rankings were treated as the response variable, while THE and ARWU positions served as predictors. Model performance was evaluated using leave-one-year-out cross-validation for a median baseline, the average of THE and ARWU ranks, linear regression, Random Forest, and XGBoost. Random Forest achieved the lowest mean absolute error (MAE = 24.80 ± 3.26 ranking positions), whereas linear regression best preserved relative ordering (Spearman’s ρ = 0.794 ± 0.062). Additional analyses of the TOP-50, TOP-100, and TOP-200 subsets showed that prediction accuracy decreases as ranking depth increases. The findings demonstrate that QS, THE, and ARWU provide complementary rather than interchangeable assessments of university performance. The proposed framework is intended for comparative cross-ranking analysis rather than forecasting future university rankings.