Research On Dynamic Weighted Heuristic Trust Path Search Algorithm

Online social networks have provided a platform for users to interact with each other and share information.The authenticity of information on the Internet sometimes is difficult to identify.Obviously,information from trusted friends is more likely to be adopted than information from strangers.Therefore,trust plays an essential role in online social networks.Many successful recommendation systems consider trust relationships.It is shown that incorporating trust into recommendation systems can improve the quality and coverage of recommendations.For example,Advogato and Film Trust both use trust networks to express trust relationships between users.Social networks have few direct trust relationships,conversely,there are more indirect trust relationships.Therefore,how to quickly search and form a reliable and optimal indirect trust relationship is a crucial issue.For users without direct interaction,as long as there is at least one trust path between two users,trust inference can be carried out through trust propagation.Therefore,path length and propagation method are two important factors in the trust inference.The thesis considers the heuristic trust path search:on the one hand,the dynamic weighted heuristic trust path search is proposed.Firstly,this algorithm finds all admissible trust paths based on the six-degree space theory.Secondly,the A~*algorithm is improved to find the optimal trust path by using a dynamic weight,the number of hops,and the depth of nodes.Thirdly,the trust propagation function is used to calculate the final trust value.Compared with the other algorithms,it not only provides users with accurate and high-quality recommendations,but also improves the efficiency of recommendations.On the other hand,the anytime dynamic weighted heuristic path search algorithm is proposed.The algorithm includes two stages:the first stage is called the exploration stage,it aims to find any solution quickly;the second stage is called the update stage,which considers the error between the current cost and the current solution,and introduces a dynamic factor w to dynamically adjust the cost range to speed up search efficiency.Experiments demonstrate that the performance of this algorithm is better than other algorithms.