A method and device for sorting search results

Abstract

The invention provides a method and a device for sorting search results. The method includes: when receiving a search request including a current search term, acquiring historical click rate and impression of objects included in the search results corresponding to the current search term according to historical statistical data; acquiring preliminary estimated click rate of the objects according to the current search term and features of the corresponding objects; acquiring final estimated click rate of the objects by nonlinear fitting according to the historical click rate, the impression and the preliminary estimated click rate; sorting the objects according to the final estimated click rate. By the method and device, accurate click rate estimation is achieved effectively, further more accurate search result sorting is realized, and user requirements are met while user friendliness is realized.

Description

technical field [0001] The present application relates to the field of computer network information, in particular to a method and device for sorting search results. Background technique [0002] With the rapid development of Internet services, various information interactions through the Internet have become the most widely used today. However, when multiple services interact with applications at the same time, interactive application congestion will occur. For example, when a user wants to search for a certain product, many product lists will appear at the same time, but not all of the products presented to the user are what the user needs, and many of them are even irrelevant to the search content. The ordering of the results is not rationalized or precise enough. Therefore, it not only brings a lot of inconvenience to user applications, but also brings a waste of network resources. This also reflects the importance of reasonable and precise sorting. [0003] The meth...

AI Technical Summary

Problems solved by technology

Although the method of sorting search results based on the click-through rate estimated by this model reduces some unnecessary waste of resources, its accuracy is not high and sometimes relevant important information is missed

[0004] For example, in the above methods, only the text features and relevance features of the objects that the user has clicked are paid attention to, without considering other important features of the objects themselves that also play a key role in the user's click behavior, such as visual perception of pictures and labels. Information that is difficult to be characterized, so it cannot accurately reflect the user's click behavior, that is, the prediction accuracy needs to be improved

In addition, because we only focus on the objects that users have clicked, some objects that have not been clicked by users temporarily or have a low amount of clicks are ignored, and such objects are likely to contain the information that a specific user currently wants to search for

When users search for such objects, they cannot be searched due to the small amount of clicks, which leads to retention and waste of resource information

In addition, for an object with a large number of user search behaviors and a small number of user click behaviors, because the user clicks less on it, the object may be ignored and not presented to the user

Therefore, the method in the prior art does not fully and rationally utilize the historical data information in the network log

Images