Reliable data-based recommendation service method

Abstract

The invention discloses a reliable data-based recommendation service method and solves the problems of low data reliability and low prediction accuracy in the prior art. Through the method, a more accurate item can be recommended to a user; and original training set data is hidden by adopting a method for disturbing the training set data, so that the problem of user information leakage in the prior art is solved, the user information leakage is avoided, and user information is protected. The method comprises the steps of (1) extracting a prediction data set; (2) disturbing the training set data; (3) calculating a reliable score of the item by the user in a training set and three sub-training sets; (4) predicting an item score; (5) assessing the predicted score; and (6) obtaining the recommended item. According to the method, the prediction accuracy is improved and the user information is protected.

Description

technical field [0001] The invention belongs to the field of computer technology, and further relates to a method for recommending services based on credible data in the field of network application layer communication technology and mass data analysis technology. The invention quantifies the credibility of the network commodity data, and predicts the score of the network commodity on the data, so as to provide personalized commodity recommendation service for network users and realize the protection of user information at the same time. Background technique [0002] The recommendation service based on credible data refers to the use of highly credible data to predict the ratings of products that users may be interested in after quantifying the credibility of network data, so as to provide users with more accurate product recommendation services. . At present, the common method of information recommendation is to use the similarity of users to provide users with more accura...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that only the support of the user is considered, but the credibility of the user's rating data is not considered, which leads to the problem of low prediction accuracy

The disadvantage of this method is that it does not consider the credibility of the user's historical rating data in the user list, nor does it consider the protection of user information, which leads to the leakage of user information

The disadvantage of this method is that the credibility of the historical scoring data of these users is not considered when dynamically selecting different numbers of neighboring users, which leads to the low reliability of the data of the selected neighboring users and the occurrence of prediction accuracy. low problem

Images