Personalized recommendation method with socialization information fused

Abstract

The present invention provides a personalized recommendation method with socialization information fused. The method comprises the following steps: S1, constructing a user- user trust matrix; S2, constructing a project- project tag similarity matrix; S3, constructing and training a model; and S4, predicting a preference of a user for an unknown project. The method provided by the present invention mainly has the following advantages: 1) a sorting learning method in the information retrieval field is applied to Top- K recommendation, so that the sorting problem in the recommendation system is effectively solved, and the defect that the conventional score-based prediction method can not perform Top-K recommendation is overcome; and 2) socialization information, i.e. user social information and project tag information, is fused into the model based on sorting learning, so that accuracy of recommendation results is improved.

Description

technical field [0001] The invention relates to the fields of personalized recommendation, ranking learning and social network, in particular to a personalized recommendation method integrating social information. Background technique [0002] With the rapid development of Internet technology, especially e-commerce, the growth rate of data in the Internet far exceeds the speed of human reception, and the problem of information overload is becoming more and more serious. Information filtering technology, which helps us to filter useful data from massive data, is becoming more and more important. Personalized recommendation technology is an ideal method to find data of interest to users from large-scale data according to user preferences. [0003] At present, the application of personalized recommendation is mainly divided into two categories. The first category is the rating prediction problem, which is to predict the rating of an unknown item by giving a user's historical r...

AI Technical Summary

Problems solved by technology

The second category is the Top-K recommendation problem. The Top-K recommendation is dedicated to recommending the K items that users are most likely to like.

[0007] The Point-wise method is still a score prediction-oriented model without considering the characteristics of sorting; the Pair-wise method needs to consider the partial order relationship between all items, and the complexity of model training is too high; although the ListRank-MF method takes into account Optimizing the sorting of the entire recommendation list can solve the problem of item sorting to a certain extent, but because too little information is incorporated into the model and the impact of social information on users and items is not considered, it is still difficult to avoid "data sparsity" and The "cold start" problem has great limitations in practical applications

Images