Improving Top-N Recommendation Performance Using Missing Data

Algorithm 1

The neighbor-based sampling scheme.

Input: the rating matrix , the random ratio , the neighbor size

Output: the neighbor-based sampling matrix

(1)  for each user   do

(2)  Find : the top- most similarity users of ;

(3)  Find : the item set, in which items have not been rated by user ;

(4)  Find : the candidate item set, a sub set of , in which items have not

    been rated by all users in ;

(5)  Random select percentage of items in into ;

(6)  end for