Differentially Private Recommendation System Based on Community Detection in Social Network Applications

Algorithm 3

Differential privacy based on community detection.

Input: Divided communities in G,;

Known labels of users;

Output: Unknown noise labels of users;

Unknown noise labels of users, users in community i.

Unknown noise labels of users, users in independent community i.

for do

for do

end for

end for

Find the max probability in Community i,

Get