CGPP-POI: A Recommendation Model Based on Privacy Protection

Algorithm 2

The recommendation of CGPP-POI.

Input:;

   ;

   ;

   ;

Output: Recommendation list ;

1: To choose the extend range

2: 

3: fordo

4:   

5: end for

6: 

7: 

8: Initialize each node to a separated community

9: repeat

10:  fordo

11:   fordo

12:    Remove i from its community, place to ’s community

13:    Compute the composite modularity gain

14:   end for

15:   Choose with maximum positive gain (if exists) and move to ’s community

16:   Otherwise, stays in its community

17:  end for

18: until No further improvement in modularity

19: Get , which are also independent subgraphs

20: if the number of users in then

21:  

22:  Sort the recommended label for each user, and get recommendation list

23: else

24:  Sort by number of check-ins based on history. And take the first and build the recommendation list .

25:  end if

26: fordo

27:  fordo

28:   , to choose n , and is obtained.

29:  end for

30: end for