Hybrid pushing method and device for diversified information

Abstract

The invention discloses a hybrid pushing method and device for diversified information, and relates to the field of electronic information. The method comprises the following steps: for each type of information contained in the diversified information, determining an original score of each information item corresponding to the information type according to the type grading rule corresponding to the information type; for each information item corresponding to each information type, performing normalization processing on the original score of the information item according to a preset normalization processing rule to obtain a normalization score; calculating a type preference score corresponding to the information item according to a preset type preference rule, and determining an actual score of the information item according to the normalized score and the type preference score; carrying out the hybrid sequencing according to the actual score of each information item corresponding to each information type contained in the diversified information, and carrying out the hybrid pushing according to the sorting result. According to the method, the information items of different types are uniformly graded in combination with a plurality of factors, and the information items can be reasonably sorted and pushed.

Description

technical field [0001] The present invention relates to the field of electronic information, in particular to a mixed push method and device for diversified information. Background technique [0002] In recent years, O2O, that is, Online To Offline (online to offline / online to offline) technology has developed rapidly, which can perfectly combine online and offline. Users' needs are diversified, such as shopping, reading, searching for stores, etc. Therefore, users' online demands will also be diversified. For example, for APP, the corresponding information types of information items include: commodity type, article type And store types, etc., in order to meet the needs of users, it is necessary to sort information items of different information types, determine reasonable recommendation results and push them to users. [0003] In the first scenario, the recommendation result is determined according to the purchase rate of the product. For the case where the article contain...

AI Technical Summary

Problems solved by technology

[0005] However, in the process of implementing the present invention, the inventor found that the above-mentioned methods in the prior art have at least the following problems: First, in the first application scenario, the probability of an article guiding a user to successfully purchase a product is often small, compared with For the products included in the article, users are more likely to purchase individual products. The disadvantage of this method is that individual products will get a greater exposure rate due to the high purchase rate, and the article requires intervention to ensure the exposure rate of the article; secondly , in the second scenario, it is unreasonable to use click-through rate alone as the sorting factor of recommendation results because information items of different information categories are not the same to attract users

To sum up, the recommendation results obtained by using single-dimensional data as the ranking factor of the recommendation results are often unreasonable, and there is no technical solution that can well solve the above problems in the prior art

Images