Modified Friend Recommending Method Based on Link Prediction
-
-
Abstract
In order to select user who has the highest similarity with target user as a friend from the mass of the users, a modified friend recommending method based on link prediction is proposed in this paper. Firstly, through the links relationship, the Top-N users who have the highest similarity with the target user are recommended. Secondly, calculating the tag similarity and the common friends similarity between the target user and recommended Top-N users. Thirdly, giving appropriate weight to them, and calculate comprehensive similarity between target users and the recommend Top-N users. Finally, sorting users based on comprehensive similarity, and recommending the users who have the highest similarity to the target users. The experimental results show that the algorithm in this paper is higher than the single link prediction by 8% on accuracy.
-
-