Advertisement delivery method and advertisement delivery device

Abstract

The embodiment of the invention discloses an advertisement delivery method and an advertisement delivery device. The method comprises steps: advertisement click rate characteristic parameters are acquired from a group user history search record; the advertisement click rate characteristic parameters are classified to obtain multiple feature sub groups; factorization is carried out on the multiple feature sub groups to acquire a sub group factorization model; the sub group factorization model is subjected to iteration to acquire a predicted advertisement click rate; and according to the predicted advertisement click rate, advertisement delivery is carried out. The embodiment of the invention reduces the training calculation amount and improves the predicted click rate. A service object is displayed according to the predicted click rate, the possibility of displaying the service object meeting user requirements is improved, multiple times of page turning operation or search in other platforms by the user is avoided, the operation convenience is improved, time consumption of the user is reduced, the user experience is improved, and the use satisfaction of the user is enhanced. Resource occupancy and resource waste by a server and a client are reduced.

Description

technical field [0001] The present invention relates to the technical field of computer processing, in particular to an advertisement placing method and device. Background technique [0002] With the development of Internet technology, obtaining product information through the Internet has become a part of people's life. [0003] In order to increase popularity and promote products, merchants often place advertisements through the Internet. There are usually two ways, one is through natural search, and the other is through competitive promotion. The first method is more competitive and the optimization cycle is longer, so many merchants will choose the second method, which is faster. [0004] Competitive promotion is a multi-party game system, including: [0005] The interests of advertisers (output-input ratio, ROI); the interests of network users / search engines (user experience, as few advertisements as possible, and as high a click-through rate as possible, CTR); the i...

AI Technical Summary

Problems solved by technology

[0007] The LR model is a simple linear model that cannot directly establish the correlation between features; it needs to process samples through time-consuming and labor-intensive feature engineering (clustering some features, etc.) to estimate the click-through rate; features directly Correlation can easily lead to data sparsity and model overfitting (data long tail), which leads to a decrease in the accuracy of click-through rate estimation

[0008] Some feature correlations in the FM model are wasteful, which reduces the fitting ability of the model; the calculation speed is much slower than that of the LR model

[0009] Due to the low accuracy of the click rate prediction based on the LR model, the ranking effect is poor, which will make the product information that can meet the user's needs ranked very low, making it displayed at a later position, or even unable to display; while the FM model speed Too slow for commercial use, leading to the following problems:

[0010] On the one hand, these product information that do not match the needs of users not only occupy the resources of the platform server, but also occupy the resources of the client, resulting in unnecessary resource occupation and waste

[0011] On the other hand, users need to obtain the product information they are interested in, and may need to perform multiple page-turning operations, or even go to other platforms to search again, and then search, compare, and filter massive information to obtain relevant product information, and the operation is more convenient. It is cumbersome, consumes users' time, reduces user experience, and will greatly increase the resource consumption of other platform servers and clients

Images