Advancing Information Systems for Smart Decision Making Using Machine Learning-Based Weather Prediction

Accurate weather forecasting plays a crucial role in decision-making processes across various sectors. Traditional meteorological models often struggle with challenges such as missing data, insufficient spatial resolution, and inadequate assimilation of observational data. This study explores the integration of machine learning into weather prediction to enhance information systems for smart decision-making. The primary objective is to develop a robust machine learning-based approach for weather temperature prediction using open weather data. This research addresses missing data challenges, optimizes predictive accuracy, and establishes new methodologies applicable to Kazakhstan’s meteorological systems. The study utilizes the rp5 weather database, which provides global weather data collected every three hours. Several machine learning models, including XGBoost, CatBoost, Linear Regression, and Bayesian Ridge, were applied using the scikit-learn library. Model evaluation was conducted based on mean squared error (MSE) and feature importance analysis, including SHAP values. Among all tested models, CatBoost demonstrated superior predictive performance with an MSE of 0.240654. Further optimization using Grid Search Cross-Validation indicated that the default hyperparameters were optimal. Feature importance analysis identified key variables affecting temperature prediction, including dew point, humidity, and atmospheric pressure at sea level. The relevance of this research is underscored by the increasing need for accurate weather predictions in industries such as agriculture, energy, and disaster management. The integration of machine learning enhances traditional forecasting methods, making them more adaptable to local climate conditions. These findings contribute to the advancement of meteorological information systems, providing actionable insights that can be utilized for optimizing weather-dependent operations in Kazakhstan and beyond.