Toward a user-oriented recommendation system for real estate websites Xiaofang Yuan, Ji-Hyun Lee n , Sun-Joong Kim, Yoon-Hyun Kim Graduate School of Culture Technology, KAIST, Daejeon 305-701, Republic of Korea article info Article history: Received 30 July 2012 Received in revised form 16 August 2012 Accepted 17 August 2012 Available online 11 September 2012 Keywords: Home buyer Real estate website Housing search behavior Case-based recommendation system Ontology abstract The Internet has become a significant transaction platform for the real estate industry. However, use of the Internet does not benefit homebuyers in terms of search time, flexibility, and intuitive results. While it does encourage buyers to search more intensively, and discover and visit more properties, it also wastes more time and energy. To improve the efficiency of real estate searches, we developed an online homebuyer’s search program, based on an investigation of search behaviors, and implement a user-oriented recommendation system for real estate websites via a combination of case-based reasoning and an ontological structure. An ontological structure is employed to improve information management efficiency while case- based reasoning improves recommendation accuracy. A user test demonstrates the effectiveness of the proposed system and validates the findings of this study. The limitations of the current study are also discussed for future research and applications. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction With the prevalence of the Internet, more Internet- enabled services are becoming available. The American National Association of Realtors (NAR) On-line Technology Survey, conducted in 2011 ([34] National Association of REALTORSs Profile of Home Buyers and Sellers), indicated that when searching for a home, 88% of homebuyers chose the Internet as an information source, an increase of 14% since 2010. Real estate websites have thus been used more frequently in the residential real estate industry. Never- theless, the median time that homebuyers spent on a home search was 12 weeks, the same as in 2009 ([35] National Association of REALTORSs Profile of Home Buyers and Sellers). This implies that the use of real estate websites did not improve homebuyer’s search efficiency [11,27]. In Korea, the real estate market territory of web-based real estate search tools has expanded due to advantages of information management and ease of provision to real estate customers [22,7]. In particular, technological inno- vations have resulted in new applications that allow homebuyers to select affordable price ranges, take virtual tours, obtain information about the neighborhood and surrounding environment, access comparative data of different districts, and view property values over time. While there are services that can connect multiple real estate agencies with customers and provide varied eco- nomic information for real estate, such services only provide limited knowledge for customers. For example, the existing linear search mechanism and housing search process services restrict the ability to run a multi-attribute housing search. As with other searches, intricately connected factors are used to make housing decisions [27,38]. In Korea, web-based real estate search services have been proactively developed due to the nation’s well-constructed Internet infrastructure [22,7]. Although there are many indications for the need of a multi-attribute search system, current Korean hous- ing search services still only allow single-attribute searches [7]. Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/infosys Information Systems 0306-4379/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.is.2012.08.004 n Corresponding author. Tel.: þ82 42 350 2919; fax: þ82 42 350 2910. E-mail address: jihyunlee@kaist.ac.kr (J.-H. Lee). Information Systems 38 (2013) 231–243