Method and device for story screening and accurate presentation

Abstract

The invention provides a method and a device for story screening and accurate presentation. The method for story screening and presentation comprises: establishing a user scoring model and a grading mechanism, according to user attribute and behavior information, grading the users and dividing corresponding user classes; establishing a story scoring model and a grading mechanism, according to story attribute and user operation information, scoring stories and dividing corresponding story classes; and establishing a hierarchical recommendation model, establishing different recommendation system models for users and stories in different hierarchies, used to realize high-quality story screening and accurate presentation.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a method and device for story screening and accurate presentation. Background technique [0002] There are currently three methods for screening and displaying relevant story works: one method is to screen the story works by the popularity of the author, and at the same time display the stories according to the degree of popularity; the second is to select the stories based on the quality of the text Screening, while displaying stories with high-quality content; the third method is to use the user behavior log to dig out stories that users may be interested in through the recommendation system method. [0003] However, the above methods have their own significant shortcomings in the specific implementation process. Screening stories based on the author's popularity can ensure the high quality of the content to a certain extent, but this method cannot excavate those w...

AI Technical Summary

Problems solved by technology

[0003] However, the above methods have their own significant shortcomings in the specific implementation process. Screening stories based on the author's popularity can ensure the high quality of the content to a certain extent, but this method cannot excavate those works that are not yet famous but have high quality. , and presenting purely based on popularity will prevent most stories from being shown

If we only start from the content to screen and display, although the best results can be obtained, the processing cost is too high to adapt to the rapidly increasing amount of information. Finally, although the traditional recommendation system method can make some high-quality stories display, but today's recommendation system cannot meet the needs of mining high-quality content and accurate display at the same time

Images