鄭 品杰

Abstract: Today, artificial Intelligence (AI) technology has been wildly spread to different industries such as advertising, business, and finance industry. On the other hand, there exist industries to which application of AI has not yet been enough despite its potential. Real estate industry is one of such domains. When applying AI technology to the real estate industry, various issues are expected to be solved such as predicting the price of properties, customer reception, floor plan identification, and support of property search. This paper focuses on the support of property search. The difficulty of property search compared with such items as books, movies, and news articles is that we would seldom buy property. It also means that we usually do not have enough knowledge and experience needed to explicitly express information need and to examine retrieved results. Therefore, human experts (salespersons) help such users to find properties which they consider satisfactory. If the tacit knowledge used by human experts to estimate the value of properties could be extracted, it would contribute to support of property search by ordinary users (customers). We think utilizing different individual's strength and experience is one of important topics of community-centric systems. Aiming at extracting expert knowledge about properties suitable for renovation, this paper applies List-wise rank learning to property search. Through an experiment using search log by actual salespersons in the real estate company, the possibility of learning expert knowledge is examined.