Click rate prediction model training method and device and click rate prediction method and device

Abstract

The embodiment of the invention discloses a click rate prediction model training method and device and a click rate prediction method and device. The click rate prediction model training method in the embodiment comprises the steps that original training data is acquired, wherein the original training data comprises an exposure data set of specific media content within a preset historical duration, and the exposure data set comprises first exposure data and non-first exposure data; at least part of the non-first exposure data in the original training data is replaced with corresponding first exposure data of the specific media content, and new training data is formed; and a new prediction model is constructed according to a preset algorithm and the new training data, wherein a click rate prediction model comprises the new prediction model. Through the click rate prediction model training method and device and the click rate prediction method and device, the accuracy of click rate prediction can be improved.

Description

technical field [0001] The invention relates to network technology, in particular to a method and device for training a click-through rate prediction model, and a click-through rate prediction method and device. Background technique [0002] The rise of the Internet enables people to view different media content when browsing the same page, realizing personalized display of media content. By testing the click-through rate, we can understand the media content that different users are interested in, so as to display the corresponding media content to each user more accurately, so as to increase the click-through rate of the media content, improve the delivery effect of the media content and the page visits. [0003] Media content click-through rate estimation refers to the background system adopting a preset estimation model every time a user requests a page, based on user information, information related to specific media content, and environmental information of the specific...

AI Technical Summary

Problems solved by technology

[0005] However, the training data used in the training of the prediction model adopted by the background system in the prior art often includes the data when the media content is clicked for the first time and the data when the media content is not clicked for the first time, but only the first time is used in the actual prediction. The data when the media content is clicked, which leads to the inconsistency between offline training and online estimation, resulting in deviations in the click-through rate estimation in some specific scenarios mentioned above

Images