Cold-Start Mastered: LebiD1
Lebi Jean Marc Dali, Qin Zhi-guang · 2014
Cold-Start is one of the most difficult problems faced by web companies today in the domain of recommendation system. Cold-Start problem refers to predicting the behavior of a new user/item having no history. Common algorithms used to predict users behavior fail at addressing the cold-start problem because the algorithms are based on the user history with the company. But in this paper, we successfully solve the cold-start problem by combining two knowledge bases: one pertaining to the web company and the other pertaining to the user social network database. In this paper, we explain our method in detail.