EXPRESS: An LLM-Generated Explainable Property Valuation System with Neighbor Imputation
Wei-Wei Du, Yung-Chien Wang, Wen-Chih Peng
TL;DR
EXPRESS tackles the challenge of missing values and the lack of interpretability in real estate valuation by integrating a configurable neighbor-imputation framework with an LLM-generated feature-wise explanation layer. The approach combines dynamic nearest-neighbor search, neighbor-based imputation, a LightGBM Automated Valuation Module, and prompt-based natural-language explanations to provide transparent valuations, with a Taiwan-based dataset informing design choices and regulatory alignment. Key contributions include a flexible property configuration mechanism, scenario-aware neighbor selection, and end-to-end explainability, all implemented in an open-source system. The work has practical impact for lenders and buyers by improving trust, transparency, and usability of automated valuation in real-world settings.
Abstract
The demand for property valuation has attracted significant attention from sellers, buyers, and customers applying for loans. Reviews of existing approaches have revealed shortcomings in terms of not being able to handle missing value situations, as well as lacking interpretability, which means they cannot be used in real-world applications. To address these challenges, we propose an LLM-Generated EXplainable PRopErty valuation SyStem with neighbor imputation called EXPRESS, which provides the customizable missing value imputation technique, and addresses the opaqueness of prediction by providing the feature-wise explanation generated by LLM. The dynamic nearest neighbor search finds similar properties depending on different application scenarios by property configuration set by users (e.g., house age as criteria for the house in rural areas, and locations for buildings in urban areas). Motivated by the human appraisal procedure, we generate feature-wise explanations to provide users with a more intuitive understanding of the prediction results.