AI Predicts Commercial Real Estate Prices

Real Estate Planet, a proptech company leveraging big data and AI in real estate, is setting a new standard in the commercial real estate market with the launch of its 'Store/Office AI Estimated Price' service. This service targets multi-unit commercial buildings nationwide and is implemented through its self-developed Automated Valuation Model (AVM). By combining traditional statistical methods and cutting-edge machine learning algorithms, it boasts high explanatory power beyond mere estimations and accuracy close to actual values. Securing stable coverage is key. The verification results of prediction accuracy, especially in the Seoul area, are noteworthy. After learning from over 220,000 real transaction data points up to December 2024 and calculating market prices for January 2025, when compared with actual transactions, it demonstrated excellent performance in key indicators such as a Median Absolute Percentage Error (MdAPE) of 5.80% and a Coverage Rate of 91.64%. Its reliability is high, with 90.32% of predictions having an error rate of 10% or less. Currently, this service provides market price information for approximately 570,000 multi-unit commercial buildings in Seoul and 2.4 million nationwide. The company plans to continuously learn from the latest real transaction data to further enhance accuracy. Real Estate Planet will not stop here; within the year, it plans to sequentially expand the application of its AI estimated price model to major real estate types, including apartments, multiplex/multi-household homes, land, and buildings. CEO Jung Soo-min stated, "This will be a useful reference in commercial real estate transactions," and expressed the ambition to "continuously advance accurate and reliable real estate market price estimation services." Real Estate Planet's innovation is expected to further enhance transparency and efficiency in the real estate market.