As one of major challenges, cold-start problem plagues nearly all recommendersystems. In particular, new items will be overlooked, impeding the developmentof new products online. Given limited resources, how to utilize the knowledgeof recommender systems and design efficient marketing strategy for new items isextremely important. In this paper, we convert this ticklish issue into a clearmathematical problem based on a bipartite network representation. Under themost widely used algorithm in real e-commerce recommender systems, so-calledthe item-based collaborative filtering, we show that to simply push new itemsto active users is not a good strategy. To our surprise, experiments on realrecommender systems indicate that to connect new items with some less activeusers will statistically yield better performance, namely these new items willhave more chance to appear in other users' recommendation lists. Furtheranalysis suggests that the disassortative nature of recommender systemscontributes to such observation. In a word, getting in-depth understanding onrecommender systems could pave the way for the owners to popularize theircold-start products with low costs.
展开▼
