Research On The Generation Of Trust Path

Rapid development of Internet has released people from the age that lacks information.Users frequently communicate and interact with each other via many kinds of online platforms.While users enjoy convenient and fast services,the amount of network data also shows exponential growth.In a short time,it is more and more difficult to evade false information from a large amount of complex information and obtain the information that users really need.The emergence of recommendation system alleviates this problem and its continuous development makes it an important channel for users to quickly obtain valuable information.The traditional recommendation system based on collaborative filtering will be affected by data sparsity and cold start,and the recommendation effect is not ideal.Trust-based recommendation system uses trust relationship between users as the basis for trust prediction and as a reference basis for recommendation,so that the results of recommendation system can be reasonably interpreted and have high accuracy.In recent years,the combination of trust mechanism with machine learning,knowledge discovery and data mining has become a research focus in recommendation system.Focusing on the problem of data sparsity and cold start in traditional recommendation system model,this paper applyd trust mechanism theory to recommendation system,proposing a weighted heuristic robust trust path generation algorithm.It applys heuristic search algorithm to trust network to find the solution path,and uses the obtained solution path to predict trust.Based on the theory of trust transfer,the model weights the heuristic function of classical A* algorithm to find a reasonable trust path,which makes the prediction results more realistic.This paper further studied the dynamic change of trust in the process of transmission.The intrinsic attribute of trust determines that its value will change with the objective factors such as path length and time.The decay factor is introduced to describe the process and the formula is given.The experiments demonstrate the advantages of robust solution path found by weighted heuristic function and make some analysis of the accuracy of trust prediction results.