A New Recommendation Algorithm Based on User’s Dynamic Information in Complex Social Network

Algorithm 2

The user similarity measurement based on social user dynamic information and prediction score calculation.

Input: Target User , Target Item , Nearest Neighbor

Set , Item Set , Timestamp , Decay Rate Parameter

, Regulatory Factor ;

Output: the prediction score of target user for

the item ;

Step  1. let ;

Step  2. when , go to Step  3; otherwise, go to

Step  4;

Step  3. when , get the number of user positive and

negative response and according to

formula (4) and (5); otherwise, ++, go back to Step  2;

Step  4. when , compute the ,

of user and according to (6), (7);

Step  5. get the user similarity according to formula (9);

Step  6. get the integrated user similarity by (10) or (11);

Step  7. finally, get the predication score of user u

to item according to (12).