* Corresponding author. Tel.:# 852-2766-6349; fax:# 852-2362-9362. E-mail addresses: hmroblaw@polyu.edu.hk (R. Law), hmnorman@ polyu.edu.hk (N. Au) Tourism Management 21 (2000) 241}249 Relationship modeling in tourism shopping: a decision rules induction approach Rob Law*, Norman Au Department of Hotel and Tourism Management, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, People+s Republic of China Received 5 June 1998; accepted 3 February 1999 Abstract Traditional tourism research on relationship modeling has concentrated predominantly on multivariate econometric models, univariate time-series techniques, and gravity approaches. These relationship modeling methods, although have attained a certain degree of success in the tourism paradigm, are primarily based on mathematical functions and are numeric in nature. A major drawback of these mathematical function-based modeling techniques is their inability to handle non-numeric data. This paper presents a new approach that incorporates the rough set theory to model the relations that exist among a set of mixed numeric and non-numeric tourism shopping data. The output of the rough set approach is a group of decision rules that represents the relations in a tourism shopping information system (IS). O$cially published data from the Hong Kong Tourist Association for the period 1983}1996 were used to form the decision rules and test the forecasting accuracy of these decision rules. Empirical "ndings indicated that 94.1 per cent of the testing cases were successfully forecasted and that there was no signi"cant di!erence between the forecasted values and their actual counterparts.  2000 Elsevier Science Ltd. All rights reserved. Keywords: Tourism shopping; Rough set; Information systems; Hong Kong 1. Introduction Many tourists visit Hong Kong every year to experi- ence the excitement, attractive shopping items, appealing food, and unique East-meet-West culture. As the largest foreign exchange earner, the tourism industry stimulates Hong Kong's economy and therefore signi"cantly con- tributes to the economic success and prosperity of Hong Kong. Among the Hong Kong industries that bene"t directly from the local tourism industry are retailing, accommodation, catering, the arts and entertainment. As it derives its demand heavily from tourism shopping, the local retailing industry depends on a dispersed travel market for its survival. In this paper, `tourism shoppinga refers to visitors' total expenditures on non-F&B items in Hong Kong. These non-F&B items could have been consumed locally in Hong Kong, or could have been taken abroad. Table 1, which is based on data from A Statistical Review of Tourism, published annually by the Hong Kong Tourist Association for the 14-year period 1983 to 1996 (Hong Kong Tourist Association, 1984}1997), shows an expanding trendline for Hong Kong's tourism industry. Furthermore, shopping expenditures and receipts from tourists indicate a remarkable increase in demand for tourism shopping in Hong Kong over the period. Shopping appears to be the most popular tourism activity in many destinations (Kent, Shock & Snow, 1983; Keowin, 1989). Di Matteo and Di Matteo (1996) claimed that shopping had become a major leisure activ- ity in part due to the more leisure oriented setting for shopping. Timothy and Butler (1995) argued that the desire and necessity for shopping could motivate a tour- ist to travel. This argument was based on the elements of relaxation, #eeing from mundane routine, and challenge that are associated with shopping. Well-known as a `shopping Paradisea, Hong Kong features global shopping, which appeals to most, if not all, tourists. Heung and Qu (1997) stated the sole purpose of visiting Hong Kong, for some tourists, was to experi- ence shopping. Hence, other than being the largest share of total tourist spending, shopping actually o!ers a real 0261-5177/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 1 - 5 1 7 7 ( 9 9 ) 0 0 0 5 6 - 4