Rresearch On Bidding Strategies Of Keywords Auction For Single Search Engine

Since the appearance of search engine, it has swept around the world with its elanof momentum. The amount of information on the Internet is growing rapidly, it hasbecome the first problem for the Internet users to solve that how to obtain the usefulinformation accurately in vast amounts of information. So the search engine has becomethe necessary network tool to Internet users. Since search engine service providers andthe advertisers both realize the potential business opportunities, keyword advertisingbusiness on the search engine has developed rapidly. So far, the keyword advertisinghas become one kind of the indispensable e-commerce marketing tool to enterprises.But in the keyword bidding process, it is not the higher of the price that theadvertiser bids on the keyword, the more profits the advertiser can obtain. Keywordauction has been in a dynamic changing environment, so how to effectively takeadvantage of the advertising expenses in the keyword advertisement is an importantissue that all of the advertisers concern, it’s also the main content of the paper.On the basis of online auction theory, this paper raises the keyword advertisingbidding auction mechanism, and uses Bayesian network prediction model to providebidding strategy for advertisers. First, based on the process of the keyword auction, thispaper extracts the variables involved in the model and builds the structure model ofBayesian network on the basis of the keyword advertising bidding process and therelationship between the variables. Then, this paper determines the parameter model byusing history data of the keyword advertisement that advertiser advertising on thesearch engine. According to whether the effect of the ads’ clicks in last stage on the ads’rank in current stage is considered in the model, this paper builds dynamic and staticBayesian network prediction model respectively. Experimental results show that thepredictive effect of the dynamic Bayesian networks is more accurate than that of thestatic Bayesian network. At the same time, this paper builds a decision-making modelbased on Bayesian network, which can help advertiser find out specific bid to achievethe highest profit in the condition of no budget constraint and according to this model,advertisers are provided with bidding strategy in the condition of budget constraints.