DEVELOPMENT OF AN INTELLIGENT SYSTEM FOR HOUSING PRICE PREDICTION BASED ON DIGITAL PLATFORM DATA

This thesis presents an OAK-oriented approach to the development of an intelligent system for housing price prediction using data generated by digital real estate platforms. Online listings provide scalable market information on price, location, total area, number of rooms, building condition, and infrastructure. To substantiate the proposed system, the study compares Linear Regression, Random Forest, and Gradient Boosting models. The comparative results show that ensemble methods provide more accurate and stable estimates than the baseline linear model, while feature analysis confirms the dominant role of location, total area, and room count. The proposed system integrates data collection, preprocessing, feature engineering, validation, and automated price estimation into a unified workflow. The results indicate that digital platform data can improve objectivity and practical decision support in real estate analytics.