Blockchain-Based Rating Incentive Mechanism For Recommender Systems:A Design And Application Research

The sparsity of user ratings is an important problem in the field of recommender system.Most of the existing research cannot fundamentally solve this problem.By motivating users to provide ratings,we can effectively improve the rating supply of recommender systems,so as to solve this problem fundamentally.Therefore,this paper will study the design and application of rating incentive mechanism of recommender systems.Considering that the existing undifferentiated rating incentive mechanism is difficult to obtain high-quality user ratings,this paper firstly improves it.By measuring users' rating contribution,this paper designs a set of differentiated rating incentive mechanism based on users' rating contribution,to guide users to provide high-quality ratings.The validity of the incentive mechanism is verified by the simulation experiments on the data sets of the Epinions.Then,based on the present-biased preference theory,this paper characterizes users' utility function under different incentive mechanisms,and compares the incentive effects of the two incentive mechanisms through simulation experiments.Next,aiming at the problem of distrust in the implementation process of the incentive mechanism,this paper further designs a blockchain-based system framework to implement the incentive mechanism.Finally,by comprehensively considering the impact of the type of incentive mechanisms and the use of blockchain on the profits of recommender systems,this paper establishes corresponding profit models,and compares the optimal profit of different incentive mechanisms,as well as discussing selection conditions of different incentive mechanisms.In this paper,we find that:(1)the differentiated rating incentive mechanism can improve the quality of user ratings,but its incentive effect may decline,depending on the type of users' risk preference;(2)the use of blockchain can help eliminate users' distrust in the implementation process of the differentiated incentive mechanism;(3)when adopting the differentiated rating incentive mechanism,the use of blockchain will help to improve the optimal profits of recommender systems under certain conditions;(4)when adopting the undifferentiated rating incentive mechanism,the use of blockchain will reduce the optimal profits of recommender systems.To summarize,the rating incentive mechanism designed in this paper can be used as a supplement to the existing rating incentive mechanism.At the same time,enterprises can use the results of this paper to choose the best rating incentive mechanism for recommender systems.