Інформаційна система прогнозування вартості комерційних та житлових приміщень на ринку нерухомості

Abstract

The development of a software module for the automated valuation of commercial and residential real estate based on the Random Forest machine learning method is presented. Random Forest enables a detailed analysis of a large volume of historical real estate data, including property characteristics, macroeconomic indicators, and local infrastructure. Data preprocessing and anomaly detection were performed using the Isolation Forest method. During the development process, model accuracy was compared, and key components influencing pricing were visualized.