Concept of a Multi-Agent System for Assisting in Real Estate Appraisals

摘要

The general architecture of a multi-agent system for real estate appraisal (MAREA) is presented in the paper. The appraisal data warehouse is filled with data drawn from source cadastral databases. Data driven appraisal models are created using different machine learning algorithms. Architecture of the MAREA system in the JADE platform was also proposed. Several experiments aimed to assess the usefulness of different machine learning algorithms for the system were conducted using the KEEL tool.