Research And Implementation Of Online Service Recommendation Methods Oriented To Inconsistent User Evaluation Criteria

The popularization and rapid development of the Internet have brought people into a new era of data and information.Internet information data has grown exponentially,which has greatly increased the difficulty for people to obtain the information they need.This is information overload.At present,one way to effectively solve information overload is to recommend systems.The recommendation system can predict the user's potential interest service based on the user's ratings and other data and recommend it to the user.It plays an extremely important role in helping the user to obtain network information.Currently,the recommendation method assumes that all users agree on service evaluation criteria and they evaluate the service according to consistent evaluation criteria.However,because of the influence of consumer background,psychology,hobbies and other factors,the user's rating preference for the service is not consistent,and the evaluation criteria cannot be consistent.If the inconsistency of the user evaluation criteria is not taken into account when recommending the online service,it will lead to incomparable prediction scores of the recommended service to the user,thus making the recommendation result misleading.In order to solve the above problems that experts and scholars have neglected in the research on the recommendation system,this paper proposes an online service recommendation method for inconsistent user evaluation criteria.According to the user's preference relationship to the online service rather than the traditional rating,the user is provided with the recommendation service.First,the user-service rating matrix is ??used to obtain the preference relationship of each user to the online service,and the preference relationship between the users is measured by the user's preference relationship.And use this to score the service forecasting,and finally according to the results of the ranking of the prediction scores of this method as the final recommendation results.The method can solve the inconsistency of the userevaluation criteria and the inconsistency of the evaluation criteria,resulting in the lack of comparability of the score.The main work of this article includes: Because the user evaluation criteria are inconsistent,different users' ratings of the same service cannot be directly compared.Therefore,this paper expresses the user's score on the service as the user's preference for the service,so as to solve the problem that the score does not have comparability.According to the user preference information,define a similarity calculation method between users.And based on similarity,service score prediction and service recommendation are made.Experiments demonstrate the advantages and effectiveness of online service recommendation methods for inconsistent user evaluation criteria.According to the algorithm design and reality online service recommendation system.