Research Article
An Efficient Recommendation Algorithm Based on Heterogeneous Information Network
| Input: heterogeneous information network, evaluation matrix R; | | Learning rate adjustment parameter regular parameter | | Output: Hidden Factors for Users and Entities | (1) | for r = 1 to R do | (2) | forto N do | (3) | Use Equation (8) to obtain the representation of neighbor nodes | (4) | end for | (5) | end for | (6) | initialize | (7) | initialize with standard normal distribution | (8) | while not convergent do | (9) | Randomly select a triple | (10) | updateby MF; | (11) | for l = 1 todo | (12) | calculation | (13) | Update | (14) | end for | (15) | Update | (16) | for l = 1 todo | (17) | calculation | (18) | Update | (19) | end for | (20) | update | (21) | end while | (22) | return; |
|