Research Article
A Trust Inference Method Employing Combinatorial Strategies
Algorithm 1
A trust inference method employing combinatorial strategies.
Input: Trust network , source user , target user , target user’s neighbor set | Output: Predicted trust value | 1: Initialize the capacity of the memory pool as | 2: Initialize the weights of the network | 3: Initialize the target- network | 4: repeat | 5: repeat | 6: Selecting the next user from the action set using the policy | 7: Get the reward value for action and the next state | 8: Store into | 9: Get samples from | 10: set | | 11: Gradient descent update for | 12: Reset every steps | 13: until | 14: Path strength Minimum value of current path | 15: ifcurrent maximum path strength then | 16: | 17: end if | 18: until All neighbors of the target user are accessed | 19: Calculate the trust prediction value using Equation (6) |
|