IMPROVING THE ACCURACY OF PRICE MONITORING FOR CONSTRUCTION MATERIALS AND THE ROLE OF INFORMATION TECHNOLOGY IN THIS PROCESS

Abstract

This article addresses the problem of improving the accuracy of price monitoring for construction materials under conditions of data fragmentation, frequent market fluctuations and limited resources. The study justifies a comprehensive approach that combines optimized sampling of price offers, the implementation of regional price indices, automated data collection, the application of machine learning algorithms and a multi-level data quality verification system. The methodology proposed in this paper ensures the reliability, efficiency and adaptability of price monitoring in the face of dynamic market conditions. A key contribution of the study is the development of a methodology for forming regional price indices based on representative materials and regional market characteristics. The approach includes collecting and classifying data by material groups, calculating weighted regional averages and integrating the results into cost estimation software. The study also highlights the importance of predictive modeling using machine learning to capture non-linear price dynamics and forecast short-term trends with high precision. These models are especially useful when planning procurement strategies under uncertainty and during periods of rapid inflation. The proposed solution addresses challenges such as inconsistent data sources, lack of standardization, limited frequency of updates, and insufficient real-time responsiveness of traditional monitoring systems. The automation of quality control processes through syntax checking, anomaly detection, source consistency validation, and relevance scoring plays a central role in improving data integrity. These mechanisms significantly reduce the influence of erroneous entries and enhance the credibility of price analyses. The findings support the need for integrated digital monitoring systems capable of dynamically adjusting to market conditions while maintaining methodological transparency. Further research should focus on refining predictive models, extending the range of monitored materials and exploring scalable tools for real-time data acquisition and verification.