Item recommendation method for combining user comment content and grades

Abstract

The invention discloses an item recommendation method for combining user comment content and grades. A model for combining the user comment content with user grades is provided for solving the cold start problem and the poor interpretability problem in a recommendation system. By means of the rich information contained in comments, the prediction accuracy is greatly improved; particularly, when data is quite sparse, the cold start problem and the poor interpretability problem can be well solved. The method mainly considers that descriptions of item characteristics are contained in user comment information, and potential characteristics in numerical value grades correspond to item characteristics in comment information through a mapping function. A model can be well built for user preferences, and therefore prediction and recommendation can be well carried out even if data is quite sparse.

Description

technical field [0001] The invention relates to an item recommendation method, in particular to an item recommendation method combining user comment content and ratings. Background technique [0002] Since Web2.0, the amount of information on the Internet has increased exponentially. Faced with massive data, users obviously feel that it is difficult to find the content they are really interested in, so the recommendation system has become more and more indispensable. . We can rely on the recommendation system to find our favorite songs from the millions of songs collected in music streaming service platforms (such as Netease Cloud Music); we can also rely on the recommendation system to find our favorite songs from news websites (such as Tencent News). news of interest. Shopping sites such as Taobao use recommender systems to recommend items that users might like. [0003] Although recommender systems perform well in many areas of the Internet, existing methods still have...

AI Technical Summary

Problems solved by technology

[0003] Although recommender systems perform well in many areas of the Internet, there are still deficiencies in existing methods

One of the disadvantages is that most recommendation systems cannot handle the cold start problem well. The cold start problem refers to that when there are new users in the recommendation system, due to the lack of sufficient historical behavior information of the new users in the system, this leads to The recommendation system cannot make satisfactory recommendation results for new users; similarly, the recommendation system cannot recommend new items to users in a timely manner.

In early studies, researchers also realized the advantages of using text content in recommendation systems, so they did a lot of work on the combination of content-based filtering and collaborative filtering. Researchers found that there are many features that affect users' perception of item ratings, they use textual review content to learn user weight distributions on these features, however their method requires experienced experts to pre-define these features

Moreover, the features selected in this way are also very random. If the features are not well selected, the system cannot correctly learn the user's preferences.

Images