Online Service Reputation Measurement for Maximizing User Group Satisfaction Based on Double DQN
The traditional reputation measurement cannot reflect real-time changes in the quality of services and user needs in a dynamic environment.Online service reputation measurement mechanism relies on user feedback ratings and serves users,user satisfaction is affected by the mechanism.A reputation measurement mechanism that does not consider to maximize of user group satisfaction makes it difficult to attract user groups.In addition,the inconsistency of user evaluation criteria makes it difficult to directly compare the ratings of different users,which leads to non-objective reputation measurement results.To address these issues,a reputation measurement meth-od based on Double DQN for maximizing user group satisfaction with online services is proposed.The method first defines the user group satisfaction with the Kendall tau distance for the case of inconsistent user evaluation crite-ria.Secondly,the online service reputation measurement problem in a dynamic environment is modeled as an opti-mization problem to maximize user group satisfaction through the Markov decision process.Finally,the Double DQN algorithm is used to solve this optimization problem.The effectiveness of the method is verified by the exper-iments on the Movielens dataset.The experimental results show that the method can give an ideal reputation deci-sion-making strategy,which maximizes user group satisfaction.
online servicereputation measurementuser preferenceKendall tau distancedeep reinforcement learning
NETL
NSTL
万方数据
