Abstract: In social media,users update their statues in real time,and the links between them change quickly,which poses a huge challenge for link prediction in such social networks．Traditional link prediction algorithms are usually efficient in specific situations,but for other situations,they are not efficient．In order to deal with the deficiency of single link prediction algorithm,a Skyline query based link prediction approach in social networks was proposed．The algorithm combines multiple link prediction algorithms,assumes the values of these algorithms as a vector of the predicted link,and returns users with the Skyline points based on the calculated vector．Experiments show that the Skyline query based link prediction approach is more efficient than related researches,and can be used in real applications of link pre-dication and recommendation in social media.

Key words: Social networks,Link prediction,Skyline,Algorithm