Air Pollution Forecasting in Almaty using Ensemble Machine Learning Models

Background: Almaty, located in a mountain–valley basin, frequently experiences stagnant conditions that trap pollutants and cause sharp diurnal contrasts in air quality. Current forecasting systems either offer detailed physical realism at high computational cost or yield statistically accurate but physically inconsistent results. Urban air quality in mountain–valley cities is strongly shaped by thermal inversions and weak nocturnal ventilation that trap pollutants close to the surface. We present a hybrid physics–machine-learning framework that combines a Navier–Stokes surface-layer model with data-driven post-processing to produce short-term forecasts of wind, temperature, and particulate matter while preserving physical consistency. The approach captures diurnal ventilation patterns and the well-known negative linkage between near-surface wind and particulate loadings during wintertime inversions. Compared with purely statistical baselines, the hybrid system improves short-range forecast skill and maintains interpretability through physically grounded diagnostics. Beyond Almaty, the workflow is transferable to other mountain–valley environments and is directly actionable for early warning, traffic and heating-related emission management, and health-risk communication. By uniting physically meaningful fields with lightweight Machine Learning correction, the method offers a practical bridge between computational fluid dynamics and operational decision support for cities facing recurrent stagnation episodes. Aim: Develop and verify a method for the diagnostics and short-term forecasting of surface circulation and particle concentrations in Almaty (2024), ensuring physical consistency of fields, increased forecast accuracy on 6–24 h horizons, and interpretability of risk factors. Compared to purely statistical baselines (R² ≈ 0.55 for PM forecasts), our hybrid framework achieved a 16% gain in explained variance and reduced RMSE by 25%. This improvement was most evident during winter inversion episodes. Methods: This study introduces a hybrid modeling framework that integrates the Navier–Stokes equations with machine-learning algorithms to diagnose and forecast surface air circulation and particulate matter concentrations. The approach ensures both physical consistency and improved predictive accuracy for short-term horizons (6–24 h). The Navier–Stokes equations in the Boussinesq approximation, the energy equation, and K-closure particulate matter transport were used. The numerical solution is based on the projection method (convection—TVD/QUICK, pressure—Poisson equation). The ML module is gradient boosting and decision trees for meteorological parameters, lags, and diagnostic quantities. The 2024 data are cleaned, normalized, and visualized. Results: The hybrid model reproduces the diurnal cycle of ventilation and concentrations, especially during winter inversions. For 6 h: wind RMSE ≈ 1.2 m/s (R² ≈ 0.71), temperature RMSE ≈ 1.8 °C (R² ≈ 0.78), and particles RMSE ≈ 0.012 mg/m³ (R² ≈ 0.64). Errors are higher for 24 h. A negative relationship between wind and concentration was established: +1 m/s reduces the median by 10–15% during winter nights. Conclusions: The approach can be generalized to other mountain–valley cities beyond Almaty. Combining the physical model and ML correction improves short-term predictive ability and maintains physical consistency. The method is applicable for air quality risk assessment and decision support; further clarification of emissions and consideration of urban canyon geometry are required. The results support early-warning systems, health risk communication, and urban planning.

Dynamic Model for Budget Allocation in via Multi-Criteria Optimization

One of the most important tasks of improving the information technology infrastructure of an enterprise is to increase the efficiency of the incident management system. The relevance of this study lies in the fact that at present the work of the technical support service in the conditions of a large flow of applications accelerates violation of the deadlines for resolution established by the business. It, in turn, leads to downtime of information systems and financial losses of the enterprise. This article analyzes the feasibility of introducing a third line of technical support to increase the proportion of incidents resolved within the framework of the Service Level Agreement adopted at the enterprise. A comparative analysis of the widely used two-level model with the proposed three-level model in this work is considered, using business process model notation. The effectiveness of the model is confirmed by automated computations using metrics, by calculating the rate and satisfaction coefficients within the framework of two and three levels of the model and then comparing these indicators to establish patterns. Thus, it is possible to track how successfully and timely incidents of information systems are resolved, which in turn directly reflects the availability and correct functioning of systems and the entire company. The company's practical losses due to system downtime were calculated, as well as the resulting financial losses before and after the adopting of the three-level system, taking into account the associated costs to identify if the initiation of the model is justified and profitable. Thus, the proposed model can be adopted by organizations in order to improve the quality of services provided by the IT department, to reduce the effect and impact of incidents on the performance and availability of systems that affect the formation of financial statements

This paper presents a three-dimensional model of pollutant diffusion in the atmosphere. The model is based on the advection-diffusion equation taking into account pollution sources and decay processes of substances. The numerical solution of the equation is realized using the finite difference method on a three-dimensional grid. The model takes into account the spatial distribution of pollutant concentrations, the influence of wind and atmospheric diffusion. Key parameters include diffusion coefficient, wind speed components, intensity of pollution sources and decay coefficient of substances. Special attention is paid to the vertical distribution of pollutants, which allows for more accurate modeling of atmospheric processes. The presented model can be used to predict the spread of pollutants, assess the impact of different sources on air quality and develop strategies to reduce air pollution in urban and industrial areas.

This study applies K-means clustering to analyze player performance in competitive Pokémon TCG tournaments, categorizing players into four distinct performance groups based on metrics such as wins, losses, and ties. Using a dataset comprising over 186,000 players, the study identifies key clusters representing varying levels of success in the game. The data was preprocessed by handling missing values and standardizing features to ensure uniform contribution across metrics. MiniBatchKMeans was employed to optimize clustering for large datasets, resulting in a model that groups players into low, moderate, and high-performance categories. The clustering results provide valuable insights into the distribution of player performance and help identify trends in competitive dynamics. A Silhouette Score of 0.4582 indicates that the clustering is moderately effective, with some overlap between clusters, suggesting that further refinement may be needed. Visualizations, including scatter plots, box plots, and heatmaps, were used to interpret the cluster characteristics, showing that top-performing players cluster into smaller groups, while a large majority of players exhibit moderate performance. The findings offer important implications for both players and tournament organizers: players can refine strategies based on their cluster profiles, while organizers can use clustering insights to design more balanced and engaging tournament formats. Future research could explore alternative clustering methods and incorporate additional performance features to further optimize player segmentation and enhance tournament design.