Research On Real-time Bidding Of Online Advertisement

With the rapid development of the Internet,more and more people choose to use the Internet to obtain information,more and more people choose to use the Internet to obtain information.At present,due to the precise audience demand of advertisers and the media's pursuit of high-quality traffic premium,real-time bidding is gradually becoming the main delivery method for display advertising.In the real-time bidding system,the demand platform serving the advertiser needs to bid on every advertisement request sent from the advertising trading platform.according to the line set by the algorithm to calculate a price for bidding.The price of bidding not only affects the chances of advertisers being exposed,but also the revenue of advertisers.Therefore,designing a reasonable bidding strategy is an important part of real-time bidding.This article will study the bidding algorithm and budget management of the bidding problem in real-time bidding.The main research contents are as follows:(1)linear bidding strategy is a commonly used bidding model in the industry,because it considers the request value of advertisement through the click rate of advertisement.However,static bidding strategy is unable to adapt to a complex and intense auction environment composed of multiple parties.,and linear bidding strategy will suffer all the performance degradation in the case of low budget.In order to solve the shortcomings of linear bidding strategy,this paper proposes a dynamic bidding strategy based on proportional control.The strategy adjusts the bid dynamically by controlling the proportion of high and low quality traffic.When the proportion of low quality traffic is higher than the set threshold,the model will reduce the acquisition of low quality traffic by lowering the price.When the proportion of high quality traffic obtained is lower than the set threshold,the model will increase the probability of winning advertising auction by increasing the price,so as to increase the acquisition of high quality traffic,so as to adapt to the complex and changeable auction environment composed of multiple demand side platforms.(2)when the advertiser's budget is limited,the advertiser's return on investment can be improved by optimizing the budget allocation in each time period and spending the budget on high-quality traffic.This paper analyzes the data set of iPinYou and finds that the time difference has a certain impact on users' clicks.In budget management,the time factor,the amount of traffic and the click-through rate of traffic are combined to allocate the budget.