????? ?? ??? (Relationships among Leisure Products)

?? ??? ? 52?? ???, ???? ??? ?? ?? ?? ?? ??? ??? ??? ??? ?? ?? ???, ???? ??? ???? ???? ??? ???? ??. ????? ??? ????? ???? ??? ?? 19?? ????? ?? ?? ??? ??? ??. ????? ?? ????? ??? ???, ???? ? ?? ?? ???? ??? ?? ?? ??, ????. ????? ??? 29? ?????? ?????.
? ????? ????? ??? ????? ??? ???? ?????? ?? ????? ?? ???? ???? ?????. ????? ???? ??? ??? ???? ??, ??? ??, TV??, ????, ??? ?? ???? ?????? ????. ????? ???? ???? ??? ??????? ??? ??? ??? ??? ??? ??????? ??? ???? ???, ????? ??? ??? ???? ?? ???? ????. ? ? ??? ?? ???? ?????? ???? 4?? ??? ????? ? ??? ?? ??? ??, ?? ???? ?? ??, ??? ??? ??? ????? ??. ?? ? ???? ?? ??? ????.
??? ??? ???? ??????? ???? ?????? ???, ?????? 4???? ?????. ???? ???? ??? ?????, ??? ??? ???? ??? ???? ????. ?? ????? ??? ??? k-nearest neighbor ??? ???? ??? ????? ?? ??? ???????? ???? ??? ??? ???? ???? ???? ???. ????? ?? ???? ??? ??? ???? ??. ??? TV?? ???? ?? ????? ??? ?? ??, ??? ?? ???? ?? ???? ??? ??? ???? ???? ?? ??? ?????? ??? ? ?? ????.

Leisure market has been expanding rapidly due to 52 labor hours rule and culture of work life balance. Consumers enjoy various leisure activities and engage in the 19 activities on average. In this study, we focus on the relationships among leisure products and use 29 leisure products for empirical analysis.
Instead of using interval scale evaluation method, we forced respondents to choose all the leisure activities that they engaged in using nominal scale evaluation method. Results show that main activities include ¡°communication with friends, rest at home, watching TV, seeing movies, internet use.¡± One alternative method is to use co-occurrence matrix and find quasi-networks. But, the major drawback of this method is to consider the cases of occurring simultaneously and overlook the cases of non-consistence case and not-occurring cases. There are four combinations that two objects are selected, only one-object is selected, only the other-object is selected, and two objects are not selected. Jaccard coefficients fit this purpose for analyzing the relationships.
Using Jaccard coefficients, we use multi-dimensional methods, draw positioning map, and could divide leisure activities into 4 groups. However, those analysis do not provide enough information for relationships among leisure products. Incorporation of Jaccard coefficients and k-nearest neighbors concepts gives more information about thickness and length of network analysis. Generally, more appearance in terms of k-nearest neighbors might end up with short appearance. However, TV watching has more appearance but not short appearance. On the other hand, sport has less appearance but short appearance.