Mixed learning model-based DSP real-time bidding advertisement system

Abstract

The invention discloses a mixed learning model-based DSP real-time bidding advertisement system. The system comprises an RTB advertisement bidding engine module, an offline model training module and an online model training module, wherein the RTB advertisement bidding engine module is connected with the offline model training module and the online model training module, calls a CTR prediction engine to predict a CTR value of a corresponding advertisement in combination with a CRT model generated by one or two of the offline model training module and the online model training module, and performs corresponding advertisement release.

Description

technical field [0001] The invention belongs to a DSP real-time bidding advertisement system based on a mixed learning model, and belongs to the field of computers. Background technique [0002] Advertisers' demand for mobile programmatic buying is also increasing, which promotes the overall development of the mobile programmatic buying market. Programmatic buying is one of the important channels for developers to monetize traffic. Developers are more willing to open inventory resources. Some star apps hero App actively shift resources to programmatic buying, which promotes the market size. DSP provides advertisers with a With a comprehensive operating platform, advertisers can manage the traffic sources of multiple channels through one platform, avoiding the waste of resources caused by complicated media buying methods. DSP can use RTB to purchase advertisements on demand in real time from the advertising exchange platform. and an audience-centric buying model [0003] T...

AI Technical Summary

Problems solved by technology

[0006] 2. Anti-cheating technology

[0038] Currently, DSPs based on RTB bidding lack enough flexibility in the use of key CTR prediction models for advertising. Many of them are based on historical data and use offline machine learning techniques to train CTR prediction models. For a new DSP Or for the newly docked ADX, there is a lack of historical data to train the model. In this case, using offline historical data to train the model shows whether it is displayed or not. It is necessary to consider using online machine learning methods to train the CTR prediction model, such as online learning. The machine learning method can change the prediction and training, and the current use of the model and the lack of flexibility in the algorithm

Images