Neural network-based search sorting method, system and equipment and storage medium

Abstract

The invention provides a search sorting method, system and device based on a neural network and a storage medium, and the method comprises the steps: obtaining a plurality of different types of recall documents according to a target search statement and a target search engine; for each recalled document, extracting comprehensive characteristics between the target search statement and each recalled document; according to the comprehensive features corresponding to each recalled document and a target sorting model, a final predicted click rate corresponding to each recalled document is obtained, and the target sorting model is obtained through training of a training sample and a training label; and sorting the recalled documents according to the final predicted click rate corresponding to each recalled document. According to the method, the category relationship between the search statement and the recalled document is fully utilized, so that the problem that the click rate difference of different categories is too large is weakened, the final recommendation result is not limited to a certain category, but is distributed among different categories in a balanced manner, and the internet experience of a user is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, and in particular, to a method, system, device and storage medium for searching and sorting based on neural network. Background technique [0002] Search scenarios are an important part of Internet business scenarios. Generally, search engines need to recall based on search keywords to obtain multiple recall documents, and then sort the recall documents. The quality of the sorting results will directly affect the user's Internet usage. Feeling, when displaying recall results, whether the recall results can directly meet the needs of users or bring unexpected surprises to users is an important sorting criterion for the quality of recall sorting results. [0003] In the existing recall sorting method, take the example of searching on an APP for a car recommendation advertisement. For example, if you search for "insurance" on the APP, the search results in the existing techno...

AI Technical Summary

Problems solved by technology

[0003] In the existing recall sorting method, take a search on a car recommendation advertisement APP as an example for illustration. For example, if you search for "insurance" on this APP, the search results in the prior art are often concentrated in the category of insurance policy, but Search results for other categories related to insurance, such as services, questions and answers, and videos, are difficult to appear on the search page, so the sorting in the prior art cannot balance the differences between the recall results of different categories, and the top ones are often easily Multiple recalls of the same category while ignoring recalls from other categories, degrading the user's Internet experience

Images