Abstract—When new items are released, it is necessary to promote these items. In this situation, a recommender system specializing in new items help item providers find potential customers. This study aims to suggest a Min-Max distance-based preference boundary, and to develop a preference boundary-based recommender procedure applied for recommending new items. The basic principle is that if a new item belongs within the preference boundary of a target customer, then it is evaluated to be preferred by the customer. The new item recommendation procedure is organized in the following two phases. The first phase defines each customer’s preference boundary based on min-max distance, and the second phase decides the target customer set for recommending new items. In this research, customer’ preferences and item characteristics including new items are represented in a feature space. And the scope of boundary of the target customer’s preference is extended to those of neighbors’. Diverse algorithms are suggested for the procedure, and their effectiveness scores are measured and compared through a series of experiments with a real mobile image transaction data set. The experiment results are compared, and discussions about the results are also given with a further research opportunity.
Index Terms—Recommender Systems; Collaborative Filtering; Multimedia Content; Personalization
M. K. Jung, M. K. Jang, and H. K. Kim are with the School of Management, Kyung Hee University, 1 Hoeki-dong, Dongdaemoonku, Seoul 130-70, Korea.
J. K. Kim is with the School of Management, Kyung Hee University, 1 Hoeki-dong, Dongdaemoonku, Seoul 130-70, Korea (corresponding author to provide phone: +82-2-961-9355; fax: +82-2-961-0515; e-mail: jaek@khu.ac.kr).
[PDF]
Cite: Min Kyu Jung, Moon Kyoung Jang, Hyea Kyeong Kim, and Jae Kyeong Kim, "A New Item Recommendation Procedure Based on Min-Max Distance,"
International Journal of Social Science and Humanity vol. 1, no. 1, pp. 43-48, 2011.